《北京师范大学智慧学习研究院：2022年中小学阶段的人工智能课程_ 对政府认可人工智能课程的调研报告（英文版）（63页）.pdf》由会员分享，可在线阅读，更多相关《北京师范大学智慧学习研究院：2022年中小学阶段的人工智能课程_ 对政府认可人工智能课程的调研报告（英文版）（63页）.pdf（63页珍藏版）》请在三个皮匠报告上搜索。

1、K-12 AI curriculaA mapping of government-endorsed AI curriculaED-2022/FLI-ICT/K-12The Global Education 2030 AgendaUNESCO,as the United Nations specialized agency for education,is entrusted to lead and coordinate the Education 2030 Agenda,which is part of a global movement to eradicate poverty throug

2、h 17 Sustainable Development Goals by 2030.Education,essential to achieve all of these goals,has its own dedicated Goal 4,which aims to“ensure inclusive and equitable quality education and promote lifelong learning opportunities for all.”The Education 2030 Framework for Action provides guidance for

3、the implementation of this ambitious goal and commitments.UNESCO Education SectorEducation is UNESCOs top priority because it is a basic human right and the foundation on which to build peace and drive sustainable development.UNESCO is the United Nations specialized agency for education and the Educ

4、ation Sector provides global and regional leadership in education,strengthens national education systems and responds to contemporary global challenges through education with a special focus on gender equality and Africa.Published in 2022 by the United Nations Educational,Scientific and Cultural Org

5、anization,7,place de Fontenoy,75352 Paris 07 SP,France UNESCO 2022This document is available in Open Access under the Attribution-ShareAlike 3.0 IGO(CC-BY-SA 3.0 IGO)licence(http:/creativecommons.org/licenses/by-sa/3.0/igo).By using the content of this document,the users accept to be bound by the te

6、rms of use of the UNESCO Open Access Repository(http:/www.unesco.org/open-access/terms-use-ccbysa-en).The designations employed and the presentation of material throughout this document do not imply the expression of any opinion whatsoever on the part of UNESCO concerning the legal status of any cou

7、ntry,territory,city or area or of its authorities,or concerning the delimitation of its frontiers or boundaries.The ideas and opinions expressed in this document are those of the authors;they are not necessarily those of UNESCO and do not commit the Organization.Cover design:Marie MoncetCover credit

8、:Ryzhi/Ryzhi/SInside icon(pp.51-53):Marie MoncetCoordinator:Fengchun MiaoPrinted by UNESCOPrinted in FranceK-12 AI curricula A mapping of government-endorsed AI curricula1K-12 AI curricula A mapping of government-endorsed AI curriculaK-12 AI curricula A mapping of government-endorsed AI curricula2 A

9、cknowledgementsThe report has been produced by UNESCOs Unit for Technology and Artificial Intelligence in Education,which sits within the Futures of Learning and Innovation Team.Fengchun Miao,Chief of this Unit,conceptualized and executed the methodology for the data collection,designed and managed

10、the surveys,and led the authoring of the report.Kelly Shiohira of JET Education Services supported the data collection,analysed the survey data,carried out the curriculum mapping,and drafted the report.Appreciation is due especially to Juan David Plaza Osses and Iaroslava Kharkova,members of this Un

11、it who organized the administration of the surveys and interviews with focal experts nominated by Member States,and to fellow colleagues Glen Hertelendy and Samuel Grimonprez for coordinating the production of the report.UNESCO acknowledges with gratitude the following governmental representatives f

12、or their contributions and time during the interviews to provide more detailed information about the AI curricula of their respective countries:Noha Alomari,ICT Education Specialist from the Department of Curriculum and Learning Resources at the Qatar Ministry of Education and Higher Education,Peter

13、 Bauer,Head of Department of Informatics and Media Technology at HTBLA Leonding in Austria,Marie-Thrse Delhoune,Inspector of Secondary Education from the General Inspection Service at the Fdration Wallonie-Bruxelles in Belgium,Helder Pais,Head of the Curriculum Development Department at the Director

14、ate-General for Education from the Ministry of Education of Portugal,and Zhang Xiong,Professor from the School of Computer Science and Engineering at Beihang University in China.The report has also benefitted from information collected from interviews with the following focal persons:ShaliniKapoor,B

15、ettina Culter,Anne Forbes Joyeeta Das and Lucy Qu from IBM,Anshul Sonak and ShwetaKhurana from Intel,Ki-Sang Song from the Korea National University of Education in the Republic of Korea,AlexaJoyce and Simran Jha from Microsoft,Irene Lee and Cynthia Breazeal from MIT;Muna Al Ansari from Kuwait;Laila

16、MohammendAl Atawy from Jordan,Mohammed Jumah F.Al-Enazi from Saudi Arabia;Stefan Badza from Serbia,KyungsukChang from the Republic of Korea,Saffin Mathew from India,Marlia Neres from Portugal,Ashutosh Raina from India,Ralitsa Voynova from the Republic of Bulgaria,Isabelle Sieh from Germany,PaulaThom

17、pson from Canada,ArtashesTorosyan from Armenia,Ralitsa Voynova from the Republic of Bulgaria,and Stephan Waba from Austria.Thanks are given to Patrick Molokwane of JET Education Services for desktop research support.Gratitude is also extended to Jenny Webster for copy-editing and proofreading the te

18、xt,and to Marie Moncet for designing the layout.Finally,UNESCO would like to thank the TAL Education Group for providing financial support to launch the project on AI and the Futures of Learning,through which this report was also made possible.K-12 AI curricula A mapping of government-endorsed AI cu

19、rricula3Table of ContentsAcknowledgements.2Objective and scope of the report.6Scope of the mapping.6Introduction.7A primer on AI terms and technologies.8Artificial intelligence.9AI techniques.9AI technologies.10Ethical AI.10AI literacy.11Pedagogical concepts and terminologies.11Existing frameworks o

20、f reference on AI curricula.12AI Literacy:Competencies and Design Considerations.13AI4K12:Five Big Ideas and K12 AI Curriculum Guidelines.14The Machine Learning Education Framework.16Methodology .18Data collection.18Criteria for selecting government-endorsed AI curricula.18List of government-endorse

21、d AIcurricula .19Limitations to the survey analysis.20Key findings of the analysis of government-endorsed AI curricula .21Curriculum development and endorsement.21AI curriculum development and endorsement mechanisms.21Vision and motivations for developing AI curricula.22Pilot testing and evaluation

22、of AI curricula .22Example:Qatar curriculum development foundations and principles .23Curriculum integration and management.25Allocation of curriculum hours.26Essential conditions for supporting AIcurricula.27Example:The introduction of AI by the CBSE in India.28AI curriculum content.30Main categori

23、es of AI curriculum content.30Time allocations for AI curriculum categories.30Coverage of AI curriculum categories.31Example:AI curriculum content in Austria .36Learning outcomes of AI curricula.38Methodology for analysing learning outcomes.38Framework for the categorization of learning outcomes.38M

24、apping of learning outcomes by AI categories.39Example:Progression of AI learning outcomes in the Republic of Korea.45K-12 AI curricula A mapping of government-endorsed AI curricula4Curriculum implementation.46Teacher training and support.46Learning tools and environments.46Suggested pedagogies.48Ex

25、ample:Implementation of the Information Science and Technology Curriculum for Senior High Schools,China .49Key findings and recommendations .51Curriculum development and endorsement.51Curriculum integration and management.52Curriculum content and learning outcomes.52Curriculum implementation .53Conc

26、luding comment.54References.55Appendix.58Survey sent to representatives of Member States.58UNESCO mapping of government-approved AI curricula .58General Information.58AI curriculum 1.58List of figuresFigure 1.Number of AI curricula by integration type.25Figure 2.Time allocation per year of AI curric

27、ula.26Figure 3.Per cent of curricula engaging each grade level.27Figure 4.Support for implementation undertaken.28Figure 5.Thematic approach to the interdisciplinary integration of AI into the curriculum.29Figure 6.AI implementation actors and procedures.29Figure 7.Boxplot of focus areas byper cent

28、of curriculum hours.31Figure 8.Allocation of curriculum time by topic area.32Figure 9.Percentage allocations for AI foundations.33Figure 10.Percentage allocations for ethics and social impact.34Figure 11.Percentage allocations for understanding,using and developing AI.35Figure 12.Percentage allocati

29、ons by topic area.37Figure 13.Curriculum Standards,Republic of Korea.45Figure 14.Average pedagogical engagement profile.49K-12 AI curricula A mapping of government-endorsed AI curricula5List of tablesTable 1.AI Literacy Competency Framework.13Table 2.Big Idea 1:Perception concepts and learning outco

30、mes.15Table 3.The Machine Learning Education Framework,with learning outcomes and definitions.17Table 4.K12 AI curricula,endorsed and implemented by governments.19Table 5.Governmental K12 AI curricula in development.20Table 6.Non-governmental AI curricula included in the study as benchmarks.20Table

31、7.Essential conditions for supporting AI curricula.27Table 8.AI curriculum areas.30Table 9.Curriculum engagement by topic area.31Table 10.Curriculum engagement for the AI foundations category by topic area.33Table 11.Curriculum engagement for the category ethics and social impact by topic area.34Tab

32、le 12.Curriculum engagement for the category of understanding,using and developing AI,by topic area.36Table 13.Knowledge outcome mapping.39Table 14.Skills outcome mapping.42Table 15.Values and attitudes outcome mapping.44Table 16.Suggested pedagogical approaches and specifications.48K-12 AI curricul

33、a A mapping of government-endorsed AI curricula6Objective and scope of the reportAs AI technology represents a new subject area for K12 schools worldwide,there is a lack of historical knowledge for governments,schools and teachers to draw from in defining AI competencies and designing AI curricula.T

34、his mapping exercise analyses existing AI curricula with a specific focus on the curriculum content and learning outcomes,and delineates development and validation mechanisms,curriculum alignment,the preparation of learning tools and required environments,the suggested pedagogies,and the training of

35、 teachers.Key considerations are drawn from the analysis to guide the future planning of enabling policies,the design of national curricula or institutional study programmes,and implementation strategies for AI competency development.Scope of the mappingUNESCO is investigating the current practices

36、of developing and implementing AI curricula in primary and secondary school education from a global perspective.AI curricula in this study refers to structured programmes of learning on AI-related topics that:1)are endorsed by either national or regional governments;and 2)target learners in general

37、school education from kindergarten to grade 12.This study does not cover AI curricula designed for specialized technical and vocational education and training(TVET)institutions,higher education institutions,or informal learning opportunities.K-12 AI curricula A mapping of government-endorsed AI curr

38、icula7Introduction 1 Eight countries were included in this analysis,namely China,France,Germany,India,Japan,Spain,the United Kingdom,and the United States,accounting for almost half the global population and 62per cent of GDP.A diverse range of AI technologies are currently in use internationally,an

39、d there is a growing recognition of the importance of AI in the context of labour and in terms of its impact on everyday life.There is wide consensus that AI will affect occupations at all levels of pay and education(Royal Society UK,2018,cited in the Microsoft Computer Science Framework,2021).A 201

40、8 analysis by McKinsey concluded that by 2030,70per cent of global firms are expected to adopt at least one type of AI technology.However,AI adoption will widen existing gaps between countries(Bughin et al.,2018a).Currently,in the United States,machines perform as many as 30per cent of workforce tas

41、ks(Kelly,2020).Additionally,increasing mismatches between skills being taught in schools and TVET institutions and skills needed by the job market are anticipated in correlation with higher rates of automation and AI integration(Bughin et al.,2018b).The COVID-19 pandemic has only increased the pace

42、of automation,which may result in as many as 1 in 16workers1 requiring retraining by 2030 and a further decline in the availability of middle-and low-skill jobs(Lund et al.,2021).The impact of AI technology is not limited to the workforce.AI has profound implications for culture,diversity,education,

43、scientific knowledge,and communication and information,especially insofar as they concern peace,sustainability,gender equality,and the specific challenges of Africa(COMEST,2019).These are all areas of significant interest to both international and national bodies that focus on development and policy

44、.Citizens are increasing their interactions with AI,knowingly or unknowingly.AI has been deployed to drive cars,automate customer service,identify targets for military bombs,screen applicants at national ports of entry,direct policing efforts,determine grades,select university entrants and scholarsh

45、ip recipients,and make decisions about personal finance(Engler,2021;Frantzman and Atherton,2019;Shiohira,2021).International policy guidance suggests that common areas should be pursued through different contextualized approaches such as promoting the inclusive and equitable use of AI in education;l

46、everaging AI to enhance education and learning;fostering skills development for jobs and life with AI;and safeguarding education data so that its use is ethical,transparent and auditable(UNESCO,2019a).However,currently relatively few initiatives focus on AI in K12 contexts,leading to a recent recomm

47、endation that policy-makers should provide an enabling policy environment and curricular spaces for exploring AI(Miao et al.,2021,p.34).As a leading part of the international community and conversation on technology in education,UNESCO has led a number of important developments in the AI in/for Educ

48、ation space.In 2015,the Qingdao Declaration(UNESCO,2015)included a point on exploring the potential of big data to enhance online learning,inform an understanding of student behaviour,and improve the design and delivery of online courses.The declaration urged that governments must develop policies a

49、nd systems to ensure secure,appropriate and ethical use of data,including safeguarding the privacy and confidentiality of students personally identifiable information.y The Beijing Consensus on Artificial Intelligence and Education(UNESCO,2019b)includes a series of recommendations and considerations

50、 for AI in Education.Demonstrating a strong focus on equity and inclusion,one of the recommendations in the consensus is to ensure that AI promotes high-quality education and learning opportunities for all,irrespective of gender,disability,social or economic status,ethnic or cultural background,or g

51、eographic location.y As part of the UNESCO Strategy on Technological Innovation and Education(2022-2025),in addition to an observatory and capacity building,the Organization seeks to develop standard-setting instruments and normative tools,including guidelines and frameworks,to strengthen the digita

52、l competencies(understanding,skills,and values)of teachers and learners and ensure a human-rights-based,safe,ethical,and meaningful use of K-12 AI curricula A mapping of government-endorsed AI curricula8technologies in a lifelong learning perspective(UNESCO,2021a).Transversal areas of action are the

53、 expansion of access to education,particularly for marginalized groups and individuals,and the quality of teaching and learning.y UNESCO published AI and Education:Guidance for policy-makers in April 2021 with an aim to foster AI-readiness among policy-makers(Miao et al.,2021).This report provides a

54、n orientation for its target readers on AI,including opportunities,risks,key definitions,trends in AI,implications for teaching and learning,and how education can prepare students for the AI era.It concludes with recommendations for local policy planning.y In October 2021,UNESCO launched AI and the

55、Futures of Learning,2 a project with three independent but complementary strands:(1)a report proposing recommendations on AI-enabled futures of learning;(2)guidance on ethical principles for the use of AI in education;and(3)a guiding framework on AI competencies for school students.The everyday real

56、ities of the current uses of AI and its impact on the world of work spur a sense of urgency to create international consensus on its acceptable roles in society,the expected humanistic considerations in its development and deployment,and how to equip children with the competences they will need to s

57、uccessfully navigate the existing not the future,but the existing world.The Beijing Consensus on Artificial Intelligence and Education(UNESCO,2019b)calls on all Member States to be cognizant of the emergence of a set of AI literacy skills required for effective humanmachine collaboration,without los

58、ing sight of the need for foundational skills such as literacy and numeracy.The Consensus endorses a humanistic approach to preparing all people with the appropriate values and skills needed for effective humanmachine collaboration in life,learning and work,and for sustainable development.To support

59、 the implementation of the Beijing Consensus,on 7 and 8 December 2020 UNESCO hosted the International Forum on AI and the Futures of Education:Developing Competencies for the AI Era.Participants at this event considered the competencies that citizensrequire:2 See https:/events.unesco.org/event?id=28

60、83602288“The worlds citizens need to understand what the impact of AI might be,what AI can do and what it cannot do,when AI is useful and when its use should be questioned,and how AI might be steered for the public good(Miao and Holmes,2021,p.6).The Forum emphasized the centrality of human-oriented

61、competencies,such as an understanding of the ethics of AI and its social impacts,and technology-oriented competencies,such as the skills and knowledge to use,interpret and develop AI.Subject-specific and interdisciplinary approaches to AI in education were recommended,including building on existing

62、ICT curricula and integrating analyses of the opportunities and impacts of AI into humanities,science and art courses(Miao and Holmes,2021).This report contributes further to the understanding of AI in K12 education,in particular the ways in which students are currently being prepared for life and w

63、ork in the AI era,by providing an analysis of the global landscape of government-endorsed AI curricula for grade school education and their design,content and implementation.This report is intended to inform the creation of supportive tools and frameworks,with a view to enabling the development of a

64、 guiding framework on AI competencies.It also forms one part of the work laid out in the UNESCO Strategy on Technological Innovation in Education(2022-2025)(UNESCO,2021a).A primer on AI terms and technologiesThis report engages a range of concepts and terms from both AI-specialist and education-spec

65、ialist fields.Despite the ubiquitous presence of AI in fields such as marketing,finance,and increasingly education,some decision-makers and practitioners may be unfamiliar with some of the terms used in this analysis.Similarly,it is not guaranteed that all AI practitioners and decision-makers will b

66、e aware of prominent trends in pedagogy referenced in the curricula.Therefore,this section provides a brief primer on some of the technologies,terms and pedagogies discussed in this text,to equip readers with a general understanding of each main concept.First,five terms from the field of AI are expl

67、ained in turn,and then the section on pedagogical concepts looks at several concepts including competence-based evaluation,constructivism,constructionism,and design thinking.K-12 AI curricula A mapping of government-endorsed AI curricula9Nearly all of these concepts and terms have generated at least

68、 some amount of academic debate,and have both proponents and detractors,but the purpose of this report is not to delve deeply into conflicting viewpoints.This should not be taken as an exhaustive exploration.Artificial intelligenceThe term artificial intelligence was coined in 1956 when Marvin Minsk

69、y and John McCarthy hosted the Dartmouth Summer Research Project on Artificial Intelligence(COMEST,2019;Haenlein and Kaplan,2019).AI has gained popularity owing to the rise of big data and the exponential growth of computing power(Haenlein and Kaplan,2019).The definition of AI has expanded and evolv

70、ed over time(Miao et al.,2021),and now refers to machines that imitate some features of human intelligence,such as perception,learning,reasoning,problem-solving,language interaction and creative work(COMEST,2019).The analysis in this report divides AI into two categories,AI techniques and AI technol

71、ogies.The former encompasses the methods used to build different types of AI,while the latter refers to the fields of study and products which are created by those techniques.AI techniquesThe AI techniques included in the curricula athat are analysed in this report are briefly described below:3 y Cl

72、assical AI is rule-based and uses conditional if-then statements to generate outputs.Rule-based reasoning can be used in technologies such as chatbots(e.g.If the input contains the words“what”,“price”and“?”,then return the listed product price amount).y Machine learning(ML)refers to any type of comp

73、uter program that can learn without explicit programming by accessing and processing large amounts of data.What is meant by learn is that the program can produce new outputs without being explicitly told what those outputs should be,as would be the case in classical AI.The remainder of this list is

74、comprised of some of the many different sub-categories of ML.y Supervised learning is a type of ML which is trained on known,labelled data to produce outputs.For 3 The explanations given here are derived from Miao et al.(2021),supplemented by examples and definitions from the curricula included in t

75、his report,in particular the MIT DAILy Curriculum,the AI4K12 Curriculum Framework,and the IBM Youth Challenge.4 For example,GAN technology can be used to generate images of people that do not exist(see https:/)example,a classifier is an algorithm that is designed to sort things into categories(e.g.s

76、pam or not spam)using labelled data.Decision trees are a type of classification algorithm in which a series of nodes(decision points,represented as questions)lead to branches,where the results of different response options are separated.For example,in the MIT DAILy Curriculum,which is discussed at l

77、ength later in this report,students create a decision tree to classify different types of pasta.One node might ask,Is it longer than four inches?,with spaghetti,linguine,and other long pastas on one branch to the next node and macaroni,farfalle,and other short pastas on another branch.y In unsupervi

78、sed learning,machine learning generates outputs based on clustering similarities in groups of unknown and unlabelled data.y Reinforcement learning is a type of ongoing ML which is trained to maximize a reward(for example,to return the maximum amount of currency on an investment).y Neural networks ar

79、e ML algorithms that are modelled on animal brains.They are comprised of input layers,hidden layers and output layers.In the hidden layers,data is processed in nodes based on its value and an assigned weight,and only data that passes a given threshold is allowed through.Filtered data makes its way t

80、hrough one or more hidden layers to the output layer.Learning in neural networks occurs through back propagation,an algorithm which seeks to minimize error by adjusting the weights in the hidden layer(s)of different nodes based on the correctness and influence of each nodes inputs.y Deep learning(DL

81、)refers to neural networks with multiple hidden layers.While ML in general relies on data that is structured(e.g.selected,labelled and organized into tables),DL can process unstructured data such as text and images.Neural networks and/or deep learning are used in image and speech recognition.y Gener

82、al adversarial networks(GANs)are a type of machine learning which is designed to generate new content,for example images.4 A GAN includes two deep neural networks.One of these generates content and the other evaluates it.GANs do not work particularly well with text yet.K-12 AI curricula A mapping of

83、 government-endorsed AI curricula10AI technologiesThe AI technologies included in the curricula that are analysed in this report are briefly described below:y Chatbots are computer programs designed to simulate oral and/or written conversation.5 y Computer vision is the field of AI that is concerned

84、 with deriving and using information gathered from images and visual inputs.Computer vision drives products such as automated highlight reels,self-driving cars,and quality-control tools(for the identification of defects)in manufacturing.6 y Natural Language Processing(NLP)is based on combining compu

85、ter science with computational linguistics,an interdisciplinary field for studying human language,in order to create rule-based models of human speech or text that can be used by computers.This enables computers to process and appropriately respond to human language.This technology drives computer t

86、ranslation from one language to another and the ability of technologies such as satellite navigation or digital assistants to respond to verbal commands.y Sensors are devices or systems that measure physical properties such as temperature or pressure and transmit this data to other electronics(such

87、as a computer processor).Sensors are one method of gathering the data used in AI.They are a fundamental part of the Internet of Things(IoT),systems in which actions are undertaken without human intervention based on inputs from different sensors(Mahdavinejad et al.,2018).A simple example would be an

88、 IoT irrigation system that gathers information from sensors embedded in soil and activates a watering device accordingly.7 Ethical AIAs noted,AI has a wide range of applications and many demonstrable benefits.For instance,AI provided important insights and issued alerts early in the COVID-19 pandem

89、ic.However,the use of AI also raises a number of ethical considerations.Bias can be introduced into AI through the datasets used and the choices of developers,leading to discrimination.Due to elements such as the hidden layers of some types of AI,the processes and factors in AI decision-making canno

90、t be seen,checked or redressed by humans,raising issues 5 See,for example,https:/ For more information,see https:/ See for example https:/ terms of explainability and transparency.Other challenges include balancing the use of personal data with the individual right to privacy;the security of data an

91、d potential exposure to cyber-crime;and the reinforcement of prior beliefs by AI algorithms based on user interest,which can limit peoples exposure to ideas and information and,some argue,infringe on an individuals right to freedom of expression(UNDESAetal.,2021).The First Draft of the Recommendatio

92、n on the Ethics of Artificial Intelligence(UNESCO,2020)highlights some of the key ethical challenges of AI,noting impacts on decision-making,employment and labour,social interaction,health care,education,media,freedom of expression,access to information,privacy,democracy,discrimination,and weaponiza

93、tion.The Recommendation proposes that AI should be monitored by third parties to ensure it is trustworthy and works for the good of humanity,individuals,societies,and the natural environment and its ecosystems.It sets out ten principles for ethical AI:1.Proportionality and no do harm suggests that A

94、I should have legitimate objectives and aims that are appropriate to the context,and based on rigorous scientific foundations.2.Safety and security suggests that AI should not cause damage and must protect against security risks throughout its life cycle.3.Fairness and non-discrimination suggests th

95、at AI systems should avoid bias,and that access to AI and its benefits should be shared at national,local and international levels,and be equally distributed without preference for sex;gender;language;religion;political or other opinion;national,ethnic,indigenous or social origin;sexual orientation;

96、gender identity;property;birth;disability;age;or other status.4.Sustainability suggests that the social,cultural,economic and environmental impact of AI technologies should be continuously assessed in the context of shifting goals.5.Privacy suggests that data for AI is collected,used,shared,archived

97、 and deleted in ways that protect the individual agency of data subjects,and that legitimate aims and a valid legal basis are in place for processing personal data.K-12 AI curricula A mapping of government-endorsed AI curricula116.Human oversight and determination suggests that humans or other legal

98、 entities bear responsibility for AI ethically and in law.7.Transparency and explainability suggests that people should be aware of when decisions are based on AI algorithms,and that individuals and social entities should be able to request and receive explanations for those decisions,including insi

99、ghts into factors and decision trends.Explanability is detailed further:outcomes,and the sub-processes leading to outcomes,should be understandable and traceable,appropriate to the use context.8.Responsibility and accountability reinforces the principle of human oversight and determination,and sugge

100、sts that impact assessment,monitoring,and due diligence mechanisms should be in place to ensure accountability for AI systems.Auditability8 must be ensured.9.Awareness and literacy refers to the responsibilities of governments as well as the public sector,academia and civil society to promote open a

101、nd accessible education and other initiatives focused on the intersections of AI and human rights,in order to ensure that all members of society can take informed decisions about their use of AI systems and be protected from undue influence.10.Multi-stakeholder and adaptive governance and collaborat

102、ion suggests that states should regulate data generated within and passing through their territories;that stakeholders from a broad range of civil organizations,and the public and private sector should be engaged throughout the AI life cycle;and that measures need to be adopted to allow for meaningf

103、ul intervention by marginalized groups,communities and individuals.AI literacyThe synthesis report of the UNESCO International Forum on AI and the Futures of Education under the theme of Developing Competencies for the AI Era(Miao and Holmes,2020)noted that the worlds citizens need to understand wha

104、t the impact of AI might be,what AI can and cannot do,when AI is useful,when its use should be questioned,and how it might be steered for the public good.This requires everyone to achieve some level of competency with regard to AI,including knowledge,understanding,skills,and value 8 While auditabili

105、ty is not explicitly defined in the Recommendation,this term refers to the ability of third parties to access,review,monitor and criticize algorithms(Jobinet al.,2019).9 SOLO stands for structure of observed learning outcome.orientation.Together,these might be called AI literacy.AI literacy comprise

106、s both data literacy,or the ability to understand how AI collects,cleans,manipulates,and analyses data;and algorithm literacy,or the ability to understand how AI algorithms find patterns and connections in the data,which might be used for human-machine interactions.This is an attempt to frame the sc

107、ope,structure,and main categories of the emerging field of AI literacy.This term has been used to guide the study presented in this report.Pedagogical concepts and terminologiesCompetence-based education(CBE)is a model often pursued in higher education and TVET,but it is increasingly being applied i

108、n various forms to K12 education.CBE is intended to transition education from models of fixed time and flexible learning,to flexible time and fixed learning.In CBE models,students are expected to demonstrate applied knowledge,skills and values in context through assessments,and they are given as muc

109、h additional support as needed until they meet the required benchmarks(NCLSorg,2017).At the heart of CBE is the concept of competence,a term which has evolved to describe the mobilization of knowledge,skills,attitudes and values to meet complex demands(OECD,2019,p.5).The intended competencies of a c

110、urriculum are usually expressed through learning outcomes,or what a student is expected to know,understand and be able to do upon completion of a course of study(Biggs and Collis,1982;Cedefop,2017;Kinta,2013).The terminology learning outcome is a modification of the earlier term learning objective w

111、hich ensures that the focus of the statement is on students actions or achievements rather than those of lecturers,and further are defined using measurable applications(Lopez et al.,2015;Sinha,2020).The relationship between curricula,learning outcomes and competence is complex in actualization but t

112、heoretically quite direct:a curriculum describes a set of intended learning outcomes,and assessments of students demonstrate their attainment of these outcomes through the application of knowledge,skills and attitudes/values within the domain or subject of study and,ideally,in new domains what Biggs

113、 and Colliss(1982)SOLO9 taxonomy refers to as extended abstract capacity.K-12 AI curricula A mapping of government-endorsed AI curricula12The frameworks and curricula examined for this report also reference constructivism,constructionism,computational thinking and design thinking.Constructivism(s)is

114、 a broad series of concepts in academia that apply to the ways in which knowledge is created or constructed(and at times co-constructed)by individuals through interactions with each other and their physical,cultural and institutional or systemic environments(Taber,2016).The types of constructivism o

115、ften applied in education are built largely upon the work of Piaget(1972),who outlines a theory of types and forms of learning which are and are not accessible to children at various stages of development;for example,concrete application would precede abstraction.A related concept is constructionism

116、,the philosophy that students learn best through applying knowledge to projects which hold a personal interest for them(Papert and Harel,1991).Constructionism is particularly applicable to digital curricula due to its origins in the domains of ICT and mathematics and its preoccupation with the ways

117、in which meaning is generated through the process of engaging,manipulating and changing digital artefacts(Kynigos,2015).Though constructivists and constructionists have a common base,constructionists challenge the hierarchies of knowledge that were set out by Piaget(1972),generating arguments that s

118、tudents can productively engage with more complex concepts at younger ages through the use of digital media and methods such as block-based programming(Papert,1996).Computational thinking,or the series of mental and physical processes undertaken to build a digital solution to a problem(identifying a

119、 problem,breaking it down into parts,building and assimilating solutions,and testing and refining them),is theorized to apply to an array of domains outside of computer science(Lodi and Martini,2021).The four established parts of computational thinking are sometimes cited as decomposition,abstractio

120、n,analysis and algorithms(Kush,2019).Lee et al.(2011)studied a range of computational thinking initiatives in grades K12 and determined that its processes could indeed be deployed by students of varying demographic backgrounds.They further proposed a use-modify-create learning progression model for

121、engaging with computational thinking,and noted that skilled teachers,10 See https:/ai4k12.orgdevelopmental considerations and appropriate technology were critical support mechanisms.One final tool which is presented in the context of some of the curricula included in this study is design thinking.It

122、 is presented as an analytic and creative process that engages a person in opportunities to experiment,create and prototype models,gather feedback,and redesign(Razzouk and Shute,2012).Originally developed in fields such as archeology,marketing and economics(Buchanan,1992),design thinking started eme

123、rgning within industry in the early 1990s,where it was developed as a consumer-oriented methodology to design innovative products or business models,particularly those involving technology(Hobcraft,2017).The process of design thinking includes empathizing(for example with consumers),defining a probl

124、em statement,generating ideas for solutions,and then prototyping and testing in an iterative design cycle until a desirable innovation is achieved(Hasso Plattner Institute of Design,2010).In schools,design thinking can offer a clear procedure for responding to a need for digital as well as interdisc

125、iplinary activities and competences.Existing frameworks of reference on AI curriculaThere are a few recent initiatives to map or create AI curriculum frameworks for grades K12.These include three that are detailed in this section:AI Literacy:Competencies and Design Considerations,the AI4K12:K-12 AI

126、Guidelines,10 and the Machine Learning Education Framework.This is not an exhaustive list,as a number of NGO,industry and academic organizations and/or individuals have developed AI curriculum frameworks to support their own programmes and undertakings.Some of these frameworks are in use by governme

127、nts,such as the Microsoft Computer Science Framework,and are included in the learning outcomes mapping later in this report.The three frameworks covered in this section were developed with the primary purpose of informing the development of AI curricula by a range of partners and are not linked to s

128、pecific products or courses.K-12 AI curricula A mapping of government-endorsed AI curricula13AI Literacy:Competencies and Design ConsiderationsLong and Magerko(2020)present a series of competencies and design considerations for AI literacy based on a scoping study of existing research,which sought t

129、o determine emerging themes in 1)what AIexperts believe a non-technical audience should know,and 2)common perceptions and misconceptions among learners.11 Describes the point at which AI becomes more intelligent than humans,and can be accompanied by concerns that AI would intentionally harm humans.T

130、heir scoping study reveals 17 competencies and 13design considerations.The descriptions indicate that for this proposal,competencies are universally at the lower levels of a knowledge taxonomy,largely confined to understanding,describing,and identifying.The competencies proposed by Long and Magerko

131、are outlined in Table 1.Table 1.AI Literacy Competency FrameworkCompetencyDescription/learning outcomes1.Recognizing AIDistinguish between technological artefacts that use and do not use AI.2.Understanding intelligenceCritically analyse and discuss features that make an entity intelligent.Discuss di

132、fferences between human,animal,and machine intelligence.3.InterdisciplinarityRecognize that there are many ways to think about and develop intelligent machines.Identify a variety of technologies that use AI,including technology spanning cognitive systems,robotics and ML.4.General vs narrow AIDisting

133、uish between general and narrow AI.5.5:AI strengths and weaknessesIdentify problem types that AI does/does not excel at.Determine when it is appropriate to use AI and when to leverage human skills.6.Imagine future AIImagine possible future applications of AI and consider the effects of such applicat

134、ions on the world.7.RepresentationsUnderstand what a knowledge representation is and describe some examples of knowledge representations.8.Decision-makingRecognize and describe examples of how computers reason and make decisions.9.ML stepsUnderstand the steps involved in machine learning and the pra

135、ctices and challenges that each step entails.10.Human role in AIRecognize that humans play an important role in programming,choosing models,and fine-tuning AI systems.11.Data literacyUnderstand basic data literacy concepts.12.Learning from dataRecognize that computers often learn from data(including

136、 ones own data).13.Critically interpreting dataUnderstand that data requires interpretation.Describe how the training examples provided in an initial dataset can affect the results of an algorithm.14.Action and reactionUnderstand that some AI systems have the ability to physically act on the world.T

137、his action can be directed by higher-level reasoning(e.g.walking along a planned path)or reactive impulses(e.g.jumping backwards to avoid a sensed obstacle).15.SensorsUnderstand what sensors are and that computers perceive the world using sensors.Identify sensors on a variety of devices.Recognize th

138、at different sensors support different types of representation and reasoning about the world.16.EthicsIdentify and describe different perspectives on the key ethical issues surrounding AI:privacy,employment,misinformation,singularity,11 decision-making,diversity,bias,transparency and accountability.

139、17.ProgrammabilityUnderstand that agents are programmable.Source:Long and Magerko,2020The design considerations proposed by Long and Magerko(2020)focus on pedagogical and learning methods,but also on social and interpersonal elements.Overall,they emphasize experiential learning and relevant material

140、,an appreciation for cognitive demands and child development theory,and the positioning of AI within learner contexts.The 15 specific design considerations that the researchers present are:1.Explainability:Include graphical visualizations,simulations,explanations of agents decision-making K-12 AI cu

141、rricula A mapping of government-endorsed AI curricula14processes,or interactive demonstrations in order to aid learners understanding of AI.2.Embodied interactions:Design interventions in which individuals can act as or follow the agent,as a way of making sense of the agents reasoning process.This m

142、ay involve embodied simulations of algorithms and/or hands-on physical experimentation with AI technology.3.Contextualizing data:Encourage learners to investigate who created the dataset,how the data was collected,and what the limitations of the dataset are.This may involve choosing datasets that ar

143、e relevant to learners lives,are low-dimensional and are messy(i.e.not cleaned or neatly categorizable).4.Promote transparency:Promote transparency in all aspects of AI design(i.e.eliminating black-box functionality,sharing creator intentions and funding/data sources,etc.).5.Unveil gradually:To prev

144、ent cognitive overload,give users the option to inspect and learn about different system components;explain only a few components at a time;or introduce scaffolding that fades as the user learns more about the systems operations.6.Opportunities to program:Provide ways for individuals to program and/

145、or teach AI agents.Keep coding prerequisites to a minimum by focusing on visual/auditory elements and/or incorporating strategies like Parsons problems and fill-in-the-blank code.7.Milestones:Consider how perceptions of AI are affected by developmental milestones(e.g.theory of mind development),age,

146、and prior experience with technology.8.Critical Thinking:Encourage learners to be critical consumers of AI technologies by questioning the intelligence and trustworthiness of AI applications.9.Identities,values and backgrounds:Consider how learners identities,values,and backgrounds affect their inte

147、rest in and preconceptions of AI.Learning interventions that incorporate personal identity or cultural values may encourage their interest and motivation.10.Support for parents:When designing for families,help parents scaffold their childrens AI learning experiences.11.Social interaction:Design AI l

148、earning experiences that foster social interaction and collaboration.12.Leverage learners interests:Exploit current issues,everyday experiences,or common pastimes like games or music when designing AI literacy interventions.13.Acknowledge preconceptions:Allow for the fact that learners may have poli

149、ticized or sensationalized preconceptions of AI from popular media,and consider how to respect,address,and expand on these ideas in learning interventions.14.New perspectives:Introduce perspectives that are not as well-represented in popular media(e.g.less-publicized AI subfields,balanced discussion

150、s on the dangers and benefits of AI).15.Low barrier to entry:Consider how to communicate AI concepts to learners who do not have extensive backgrounds in mathematics or computer science(e.g.by reducing the prerequisite knowledge/skills,relating AI to prior knowledge,and addressing learners insecurit

151、ies about their ability).AI4K12:Five Big Ideas and K12 AI Curriculum GuidelinesThe AI4K12 Initiative was launched by the Association for the Advancement of Artificial Intelligence(AAAI),the Computer Science Teachers Association(CSTA),and AI4All in 2018 as a joint working group that seeks to develop

152、national guidelines for teaching K12 students about AI(AAAI,2018).This group brought together academics,researchers and teachers to work towards a comprehensive AI framework based on five big ideas:1)computers perceive the world using sensors;2)agents maintain representations of the world and use th

153、em for reasoning;3)computers can learn from data;4)intelligent agents require many types of knowledge to interact naturally with humans;and,at the very centre,5)AI can impact society in both positive and negative ways.The Five Big Ideas in Artificial Intelligence poster resource has been translated

154、into 15 languages to date,12 and formed at least part of the basis for the development of curricula in multiple contexts,including several of the curricula researched for this study.The working group was convened to unpack each of these ideas into a curriculum framework divided into four parts,for g

155、rades K2;35;68;and 912.To date,curriculum guidelines for the first three big ideas have been drafted and are currently available for public comment.1312 See https:/ai4k12.org/resources/big-ideas-poster 13 See https:/ai4k12.org/gradeband-progression-chartsK-12 AI curricula A mapping of government-end

156、orsed AI curricula15In the guidelines,each big idea is subdivided into learning concepts,which are further split into concept components.For example,the learning concepts,concept components and associated learning outcomes for Big Idea 1:Perception are summarized in Table 2.Table 2.Big Idea 1:Percep

157、tion concepts and learning outcomesLearning conceptsConcept componentsLearning outcome progressionSensingLiving thingsK2:Identify human senses and sensory organs.35:Compare human and animal perception.68:Give examples of how humans combine information from multiple modalities.912:N/A Computer sensor

158、sK2:Locate and identify sensors(camera,microphone)on computers,phones,robots,and other devices.35:Illustrate how computer sensing differs from human sensing.68:Give examples of how intelligent agents combine information from multiple sensors.912:Describe the limitations and advantages of various typ

159、es of computer sensors.Digital encodingK2:N/A35:Explain how images are represented digitally in a computer.68:Explain how sounds are represented digitally in a computer.912:Explain how radar,lidar,GPS,and accelerometer data are represented.ProcessingSensing vs perceptionK2:Give examples of intellige

160、nt vs non-intelligent machines and discuss what makes a machine intelligent.35:Use a software tool such as a speech transcription or visual object recognition demo to exhibit machine perception,and explain why this is perception rather than mere sensing.68:Give examples of different types of compute

161、r perception that can extract meaning from sensory signals.912:Explain perception algorithms and how they are used in real-world applications.Feature extractionK2:Give examples of the features that one would look for if one wanted to recognize a certain class of objects or entities(e.g.cats)in an im

162、age.35:Illustrate how face detection works by extracting facial features.68:Illustrate the concept of feature extraction from images by simulating an edge detector.912:Explain how features are extracted from waveforms and images.Abstraction pipeline:languageK2:Describe the different sounds that make

163、 up ones spoken language,and for every vowel sound,give a word containing that sound.35:Illustrate how sequences of sounds can be recognized as candidate words,even if some sounds are unclear.68:Illustrate how sequences of words can be recognized as phrases,even if some of the words are unclear.912:

164、Illustrate the abstraction hierarchy for speech understanding,from waveforms to sentences.Abstraction pipeline:visionK2:Demonstrate figure/ground segmentation by identifying the foreground figures and the background in an image.35:Illustrate how the outlines of partially occluded(blocked)objects in

165、an image differ from the full shapes of the objects.68:Describe how edge detectors can be composed to form more complex feature detectors,e.g.for letters or shapes.912:Demonstrate how perceptual reasoning at a higher level of abstraction draws upon earlier,lower levels of abstraction.Domain knowledg

166、eTypes of domain knowledgeK2:Describe some things an intelligent agent must know to make sense of a question.35:Demonstrate how a text-to-speech system can resolve ambiguity using context,and how the error rate increases with ungrammatical inputs.68:Classify a given image and then describe the kinds

167、 of knowledge a computer would need in order to understand scenes of that type.912:Analyse one or more online image datasets.Describe the information that the datasets provide and how this can be used to extract domain knowledge for a computer vision system.InclusivityK2:Discuss why intelligent agen

168、ts need to understand other languages.35:Discuss how domain knowledge must be broad enough for all the groups an application is intended to serve.68:Describe how a vision system might show cultural bias if it lacked knowledge of objects not found in the culture of those who created it.912:Describe s

169、ome of the technical difficulties in making computer perception systems function well for diverse groups.Source:AI4K12(2020)Each big idea is broken down in a similar manner with a concrete learning outcome pathway from early primary school to high school.In addition to these outcomes,the curriculum

170、guidelines offer examples of the enduring knowledge that students are expected to retain,for example:Sounds are digitally encoded K-12 AI curricula A mapping of government-endorsed AI curricula16by sampling the waveform at discrete points(typically several thousand samples per second),yielding a ser

171、ies of numbers or The spoken language hierarchy is:waveforms articulatory gestures sounds morphemes words phrases sentences.Sometimes the learning outcomes and enduring knowledge are further unpacked,as was this second example:To go from noisy,ambiguous signals to meaning requires recognizing struct

172、ure and applying domain knowledge at multiple levels of abstraction.A classic example:the sentences“How to recognize speech”and“How to wreck a nice beach”are virtually identical at the waveform level.Occasionally,activities are suggested.For instance,to explain decision trees at the grade 35 level,t

173、he“guess the animal”game,troubleshooting problems,and the Pasta Land activity are recommended.The big ideas are mutually reinforcing.For example,Big Idea 3 leverages the knowledge from sensing components to facilitate a discussion of differences in how people and computers learn.Similarly,it builds

174、on the knowledge of processing components to equip learners to label a dataset for machine learning,train classifiers and engage AI concepts such as decision trees,neural networks,supervised learning,unsupervised learning,and reinforcement learning.The Machine Learning Education FrameworkAlthough it

175、 never mentions competence-based education,the Machine Learning Education Framework(Lao,2020)follows the well-known CBE framework of knowledge,skills and attitudes(which have in other contexts included items such as abilities and/or values)(Brewer and Comyn,2015;CANTA,2014;European Parliament and Co

176、uncil of the European Union,2006).CBE has in the past been criticized by some for its lack of attention to the meaning of the task for students and a reductionist view of competence which,while firmly rooted in the context of performance,is less sensitive to individual factors like prior experience

177、and the flexibility to tap into external resources,e.g.the knowledge of teammates(Rutayuga,2014).However,the gradual integration of theories such as constructivism and experiential learning(Brunner,1990;Kolb,2015;Piaget,1972;Williams,2017)has resulted in a competence-based framework that focuses on

178、head,heart and hands,in which the head represents the cognitive domain(what you know about it),the heart represents the affective domain(why it matters)and hands represent the psychomotor domain(what you can do with it)(Gazibara,2013;Singleton,2015;Sipos et al.,2008).This integration has also expand

179、ed the concept of competence to include social and emotional skills(European Parliament and Council of the European Union,2006;Mulder,2007).In addition,Lao(2020)draws on:y theories of constructionism,or the idea that learning is enhanced when undertaken through the construction of an item that has p

180、ersonal meaning for the students;y computational thinking,a proposed reframing of familiar competence concepts to apply concretely to the programming world:technical concepts,programming practices,and perspectives on an individuals relationships with technology;y a model for understanding the learni

181、ng outcomes for computational thinking lessons,divided into abstraction,or the ability to apply concepts to new use cases;automation,or utilizing a computer to increase efficiency in repeated tasks;and analysis,or reflection on a students assumptions and methods of implementation(Lee et al.,2011).y

182、Use-Modify-Create(UMC),a tiered progression often employed in computational thinking lessons,in which students first engage with existing software,and then modify it to fit new needs,and finally create new software(Lee et al.,2011).The Machine Learning Education Framework(outlined in Table 3)consist

183、s of six minimally required courses for ML-engaged citizens,and is targeted to a tinker/consumer audience(Lao,2020,p.61).In her framework,Lao makes the argument that understanding bias and the social implications of AI are fundamental requirements for all skills.K-12 AI curricula A mapping of govern

184、ment-endorsed AI curricula17Table 3.The Machine Learning Education Framework with learning outcomes and definitionsKnowledge1.General ML knowledge*Know what machine learning is(and is not).Understand the entire pipeline of the creation of ML systems.2.Knowledge of ML methodsIdentify when to use a ra

185、nge of ML methods across the breadth of the field(e.g.k-nearest neighbours,CARTs or decision trees,neural networks,ensemble methods).Understand how different methods work.3.Bias in ML systems*Understand that systems can be biased,and the different levels and ways in which bias can be introduced.4.So

186、cietal implications of AI*Understand that ML systems can have widespread positive and negative impacts.Consider the ethical,cultural and social implications of what they do.Skills1.ML problem scopingDetermine which problems can and should be solved by ML.2.ML project planningPlan a solution which is

187、 sensitive to both technical and contextual considerations.3.Creating ML artefactsUse tools to create appropriate artefacts.4.Analysis of ML design interactions and results*Describe the explicit and implicit design intentions of an ML system.Critically analyse the intentions against how the system c

188、an and should be used.5.ML advocacy*Critically discuss ML policies,products and education.6.Independent out-of-class learningStudents seek learning experiences outside the classroom.Attitudes1.InterestStudents are engaged and motivated to study the topic.2.Identity and communityStudents contribute t

189、o and learn from a community of peers and/or broader online communities who are interested in ML.3.Self-efficacyStudents are empowered to build new,meaningful things.4.PersistenceStudents continue and expand their engagement with ML.*The starred items are the six required courses outlined in the fra

190、mework.Source:Lao,2020Lao(2020)also presents a rubric for evaluating ML learning programmes against this framework,setting up the basis for a set of standards at the exit level which could be built upon.For example,the four top scores in the rubric for the four learning outcomes under Knowledge are:

191、1.General knowledge:Graduates of this course can give a precise definition of machine learning and provide a detailed description of the steps of the ML pipeline with technical and socio-ethical considerations for each step.2.Knowledge of ML methods:Graduates of this course are able to accurately di

192、scern when to use a range of machine learning methods across the breadth of the field.They are able to describe core technical concepts of these methods and comfortably use/implement them in appropriate applications.(Lao then lists her views on appropriate methods for different educational levels:y

193、High school and above:K-nearest neighbours,CART/DT,regression,convolutional neural networks;unsupervised methods such as k-means clustering,principal component analysis,GANs,and embeddings;RNNs/LSTMs;reinforcement learning;transfer learning;and ensemble methods.y Primary and middle school:Engage app

194、lications that allow students to complete specific tasks using ML,e.g.exploiting neural network and GAN applications to create art or music,or deploying reinforcement learning to play games,etc.)3.Bias:Graduates of this course are able to describe how ML systems may come to be unpredictably biased a

195、gainst specific groups throughout each step of the ML pipeline.They can critically incorporate the practices of ethical thinking and design in their own work.4.Social implications:Graduates of this course recognize that it is necessary for the creators of ML technologies to consider the societal imp

196、lications of their work.They are able to apply ethical and cultural perspectives and concepts to the analysis of ML artefacts in comprehensive,interrelational and sensitive ways(i.e.considering privacy,security,the potential for abuse,and the balance of benefits and harm;and assessing ethnographic r

197、eception and disparate impacts using tools such as stakeholder analyses,ethical matrices and modelcards).K-12 AI curricula A mapping of government-endorsed AI curricula18Methodology 14 These actors were sourced through a list of key organizations in the field of ICT in education accumulated by UNESC

198、O over the course of organizing nine editions of the Mobile Learning Week from 2011 to 2020.Data collectionTwo surveys were distributed,the first to representatives of 193 UNESCO Member States and the second to over 10,000 private-and third-sector actors.14 The surveys asked the respondents to repor

199、t on AI curricula for students in K12 general education.Appendix A provides the questions in the survey sent tothe representatives of Member States.It was modified only very slightly for the private-and third-sector actors.After the surveys were returned,the team emailed additional questions on lear

200、ning outcomes,implementation and preparation to the respondents who had indicated that they did have AI curricula at some stage of development.In addition,semi-structured key informant interviews were held with Member State representatives,non-profit leaders and developers,academics and industry pro

201、fessionals to gain further clarity on the curricula and their deployment in schools.Interviews probed the motivations for developing the AI curricula,and the reasons for their decisions around implementation methods and pedagogies.Finally,a mapping exercise was undertaken for those curricula which h

202、ad been drafted or published and were available for review.The exercise focused on the learning outcomes stated within each curriculum.Where possible,associated textbooks or materials were also reviewed to gain a further understanding of the content covered by the curriculum.Criteria for selecting g

203、overnmentendorsed AI curriculaThis study is focused on government-endorsed curricula within general K12 education.The results include only the curricula that have been,or are in the process of being,approved by national or local governments.As stated earlier,AI curricula in this study refers to stru

204、ctured programmes of learning that cover topics in the field of AI and engage with AI-related learning outcomes.Of the 193 Member States contacted through the official UNESCO channels of correspondence,a total of 51 responded,indicating at least a general interest in the topic.Among them,29 countrie

205、s and one territory completed the full survey.y Representatives from 10 countries reported no AI curricula in their country.These are:Bahrain,Canada,Colombia,Costa Rica,Estonia,Guinea,Macedonia,the Maldives,Singapore,and the Ukraine.y Representatives from 20 countries and one territory responded tha

206、t they were aware of at least one AI curriculum that was developed and endorsed by government or is under development.These are:Algeria,Armenia,Austria,Belgium,Canada(Yukon Territory),France,Germany,Jordan,Republic of Korea,Kuwait,Lao Peoples Democratic Republic,Peru,Portugal,Qatar,the Republic of B

207、ulgaria,Saudi Arabia,Serbia,Slovenia,Syria,and the UnitedArabEmirates.In addition,a total of 31 NGOs,academics and industry partners responded to the non-governmental survey and indicated that they had an AI curriculum.All Member State representatives and organizations reported having an AI curricul

208、um were contacted with follow-up questions via email and asked to provide any available curriculum documents.During the course of these follow-up emails and the curriculum mapping exercises,it was found that some of the reported curricula did not to meet the stringent criteria set out for inclusion

209、in this study.Curricula were excluded from further analysis if they did not have AI-specific learning outcomes,did not cover ordinary K12 education(e.g.were focused on TVET),were not endorsed by the government at the national or regional level,and/or did not provide enough information toanalyse.K-12

210、 AI curricula A mapping of government-endorsed AI curricula19List of governmentendorsed AIcurricula Curricula are categorized as governmental if they were provided in response to the survey distributed to UNESCO Member States and were developed by or under the directive of governmental agencies.In o

211、rder to be eligible for analysis,the survey and interview responses had to provide clear,consistent and meaningful information on the curriculum.15 This column shows the responses as they appeared in the completed surveys.After applying the selection criteria to all of the data from the governmental

212、 and non-governmental surveys,it was found that:y 11 Member States have developed,endorsed and implemented AI curricula.y The Yukon Territory of Canada has developed and implemented a curriculum entitled Applied Design,Skills,and Technologies,which has been locally rather than nationally endorsed.Ta

213、ble 4.K12 AI curricula,endorsed and implemented by governmentsCountry/regionCurriculum titleCurriculum developer15Educational levelsPrimary MiddleHighArmeniaCurriculum of ICT GovernmentXXAustriaData Science and Artificial IntelligenceFederal Ministry of Education,Science and ResearchXBelgiumIT Repos

214、itory Fdration Wallonie-Bruxelles(French-speaking Community of Belgium)XChinaAI curriculum embedded in the Information Science and Technology curriculumThe Ministry of Education of the Peoples Republic of ChinaXXXIndiaAtal Tinker Labs AI modulesAtal Tinker Labs,Atal Innovation Mission,NITI AayoagXXR

215、epublic of KoreaAI Mathematics under the Mathematics Subject Group for high schools Korea Foundation for the Advancement of Science and CreativityXAI Basics under Technology Home Economics Subject Group for high schools Korea Foundation for the Advancement of Science and CreativityXKuwaitStandards c

216、urriculumCurricula technical guidance experts and teachersXXPortugalInformation and Communication TechnologiesState school teachers of ICT and MathematicsXXXQatar Computing and Information TechnologyBinary Logic,Ministry of Education and Higher EducationXXX Computing and Information Technology(High

217、Tech Track)Binary Logic,Ministry of Education and Higher EducationXSerbia Informatics and programming Grade 8 Ministry of Education working groupX Modern technologies in gymnasiums Grade 3 and 4 Ministry of Education working groupXUnited Arab Emirates AI curriculum embedded under the Technology Subj

218、ect FrameworkMinistry of EducationXXXSource:UNESCO(2021b)In addition to the curricula being implemented,those that are in development and will likely be endorsed by the governmental agencies were also analysed.As shown in Table 5,these include an additional three AI curricula from Serbia and one eac

219、h from four more countries(Germany,Jordan,the Republic of Bulgaria and Saudi Arabia).K-12 AI curricula A mapping of government-endorsed AI curricula20Table 5.Governmental K12 AI curricula in developmentCountry/regionCurriculum titleCurriculum developerEducational levelsPrimaryMiddleHighGermany1.Iden

220、tifying and Formulating Algorithms Algorithmen erkennen und formulierenStanding Conference of the Ministers of Education and Cultural Affairs of the LnderXXXJordan2.Digital SkillsNational Center for Curriculum DevelopmentXXBulgaria3.Computer Modelling,Information Technology and InformaticsExpert gro

221、ups(academia,teachers,education experts)XXXSaudi Arabia4.Digital SkillsBinary Logic and Tatweer Co.XXXSerbia5.Technique and TechnologyMinistry of Education working groupX6.AI in gymnasiumsMinistry of Education working groupX7.AI in all high schoolsMinistry of Education working groupXSource:UNESCO(20

222、21b)Non-governmental curricula were included in the study if they covered AI learning outcomes and were at some stage of implementation in cooperation with at least one local government.However,these curricula have not been confirmed as officially endorsed by governmental agencies,and are included o

223、nly as non-governmental benchmarks.Some of these developers have undertaken other work related to the listed curricula which was also examined for the curriculum mapping.These include a curriculum adaptation entitled IBM-CBSE AI Curriculum for Grade XI&XII and the Microsoft series of textbooks calle

224、d Artificial Intelligence,Data Analytics and Machine Learning,both designed for use in India;and the Microsoft Computer Science Curriculum Toolkit.Table 6.Non-governmental AI curricula included in the study as benchmarksCountry/regionCurriculum title(s)Curriculum developerEducational levelsPrimaryMi

225、ddleHighInternational1.IBM EdTech Youth Challenge IBM XX2.AI Youth Skills Microsoft XX3.Global AI Readiness Program(High Tech Track)Intel XX4.Global AI Readiness Program(General Track)Intel XXUnited States 5.DAILy Curriculum MIT XXSource:UNESCO(2021b)Limitations to the survey analysisAs noted earlie

226、r,this analysis does not capture all of the activities related to developing AI competencies for school children,and does not even encompass all of the available information on governmental AI curricula.A wide range of curricula are outside the scope of this study.For example,three curricula were su

227、bmitted from Austria,but two operated in the TVET sector,which is not covered here.A number of for-profit providers supply training on their proprietary technology,and many AI-training programmes are offered by NGOs through non-formal learning channels like independent study;none of these were analy

228、sed.Other limitations include the following:y Some government-endorsed AI curricula might have been missed.The survey was sent out to all 193 Member States,but it is possible that some countries which do have AI curricula did not respond.y There are gaps in the data.Some of the data sought,particula

229、rly around the number of schools and learners,were not available for many curricula,which either do not track these or are not authorized to release them.y The future relevance could be questionable.The mapping is time-bound,as many curricula are still in development and may be further revised.This

230、dataset only provides a snapshot of activities in the non-governmental and private sector,and may be of limited utility in the future.K-12 AI curricula A mapping of government-endorsed AI curricula21Key findings of the analysis of government-endorsed AI curricula The following five sections present

231、the results of the analysis:1.The section on curriculum development and endorsement mechanisms addresses the mandate,motivation and means of endorsement for the AI curricula.2.The section on curriculum integration and management includes approaches to incorporating AI curricula into education system

232、s,including time allocations in terms of percentages and total hours and the preparation of essential conditions for supporting AI curricula.3.The section on curriculum content outlines the time allocations for topic areas in three main categories of content,namely AI foundations;ethics and social i

233、mpact;and understanding,using and developing AI tools.4.This section presents the learning outcomes of mapped AI curricula defined under the competence areas of knowledge,skills and values.5.The section on curriculum implementation summarizes the main strategies for teacher training and support,prep

234、aration of the required learning tools and environments,and suggested pedagogies.Curriculum development and endorsementAs listed above,14 AI curricula were endorsed and mandated for use in schools by governmental agencies in 11 countries,while 7 curricula in 5 countries are still in development.Only

235、 Serbia has both endorsed curricula and curricula still in development.A further two curricula were endorsed at the local level:the Applied Design,Skills,and Technologies curriculum from the Yukon Territory in Canada,and the MIT DAILy Curriculum in the United States.In some parts of the analysis,the

236、 four private-sector,non-governmental AI curricula were also included as benchmarks.AI curriculum development and endorsement mechanisms y Centralized government-led approachThe majority of these curricula were developed by national public agencies and endorsed through a centralized government-led a

237、pproach,at times with the participation of or in collaboration with key stakeholders.For example,in the Republic of Korea,development was undertaken by experts under a government directive;and in China,Kuwait,and the Republic of Bulgaria,development included teachers as well as academics and experts

238、.y Government-commissioned private provisionA second approach was government-commissioned private provision.In Saudi Arabia and Qatar,companies were commissioned by the government to develop the national curriculum.The Saudi Arabian representative commented that:“New technologies emerge every day,an

239、d application features are updated frequently.Therefore,we choose to work with a quality private company that has a strong reputation for ICT curriculum construction and incorporates the latest technologies and IT applications.y Government-directed decentralized approachThe third approach to develop

240、ment and endorsement was a government-directed decentralized approach.In Belgium,a parliamentary mandate created standards which were then adopted by networks of schools.These networks determine aspects such as the technologies and pedagogies to be used.A similar approach is seen in Germany,where a

241、national mandate and standards are further developed by local or provincial governments into a curriculum for implementation inschools.K-12 AI curricula A mapping of government-endorsed AI curricula22 y Private-sector-driven non-governmental AI curriculaFinally,some curricula are non-governmental an

242、d driven by private sector actors.They may be adopted as-is by schools or adapted by local experts when they develop curricula for governmental agencies.These curricula aim for a degree of flexibility so that they can be incorporated into various government frameworks and requirements,and further cu

243、stomization is also undertaken for specific country contexts.An important part of the development and endorsement of these curricula is validation,both nationally and internationally.Intels representative remarked that:“We had the curriculum extensively validated by countries.We created a global poo

244、l of validator experts and gave them the content to modify and make recommendations.When we started implementing it in 2019 and 2020,we did an extensive evaluation.But no evidence has shown that these private-sector-driven curricula have been endorsed as governmental AI curricula.In examples of this

245、 approach,industry or academic stakeholders including IBM,Intel,Microsoft and MIT have produced their own curricula and resources in consultation with experts and teachers.Vision and motivations for developing AIcurriculaThe interview respondents indicated two significant rationales for the developm

246、ent of AI curricula:to improve capacity and respond to the skills needed by the labour market,and to ensure that students graduate with the skills necessary for everyday interactions in social and political life.The extent to which these considerations were emphasized differed widely,though.For exam

247、ple,one country indicated they were not at all concerned with labour market skills,while others cited this as the primary consideration.The goal of developing skills for the labour market reflected an understanding of the shifting needs of the technology sector and the wider world of work.For most b

248、ut not all countries,this was linked to a desire to develop an internationally competitive workforce.Corporate developers also mentioned this as a strong motivation for including the development of AI training courses for students in their corporate social responsibility activities.The second ration

249、ale was associated with an understanding of AI as a driver of social as well as economic transformation,and a desire to foster general knowledge about AI and its functions and uses in society among students.Several survey respondents noted that AI is already embedded in a range of everyday interpers

250、onal interactions,and felt citizens should recognize AI in their environments,understand its benefits and potential challenges,and be empowered to advocate for safe,beneficial and transparent AI technologies.This is exemplified in the following three comments,which were contributed by the representa

251、tives from Portugal,Austria and Jordanrespectively:“We have a clear vision of the impact of technology in the future,and the need for a workforce and citizens who relate to technology in a healthy way on a daily basis.This includes the concepts,awareness,and skills to improve these areas,work with m

252、achines,and see robotics as complementary to society.This is the big picture.Artificial intelligence is seen as a transversal issue that has the potential to disruptively change key areas and concepts of life.Therefore,it is not only the expertise of specialists and developers that is of great impor

253、tance in the consideration of AI in education,but also general knowledge about the basics for all people to enable them to lead a safe and self-determined life in a world shapedbyAI.The plan is to develop a digital skills curriculum appropriate to global developments and the expected digital transfo

254、rmation,and also attain the worldwide digital competencies as relevant to our context.Pilot testing and evaluation of AI curricula The following governmental AI curricula have been implemented and evaluated:Computer Modelling,Information Technology and Informatics,Republic of Bulgaria;Information Sc

255、ience and Technology,China;MIT DAILy Curriculum;Serbias Informatics and Programming,Modern Technologies,and Technique and Technology;and the Technology Subject Framework implemented in the UAE.Additionally,the implementation of the following non-governmental AI curricula have been evaluated:IBM EdTe

256、ch Youth Challenge;both versions of the Intel AI for Youth:Global AI Readiness Program;and Microsoft AI Youth Skills.K-12 AI curricula A mapping of government-endorsed AI curricula23Some of the curricula were revised based on the evaluation,including Chinas,the UAEs,IBMs,Intels(both versions),Micros

257、ofts,and MITs.A few are still being piloted and may be further revised,namely the one from Bulgaria,the MIT DAILy Curriculum,and Serbias Technique and Technology.Common evaluation methodologies included:y Expert reviews of the curriculum.For example,in the UAE,the curriculum was shared with differen

258、t social stakeholders including academics and AI specialists.Cross-disciplinary reviews were conducted by experts in psychology and education.y Research conducted by developers.Methodologies included testing learners and carrying out interviews and surveys with teachers and representatives of region

259、al and/or national administrative departments.Information was gathered on learning outcomes,the perceived usefulness of the curriculum,and challenges in implementation.y External evaluation.Some governments commissioned external evaluations of the curriculum and/or its outcomes.For example,the Repub

260、lic of Bulgaria commissioned a national external assessment which measured the digital competencies of learners.Few of these reviews or evaluations are published.A key point surfaced by the interviews was that AI curricula should be coordinated with the mathematics curriculum and the classroom requi

261、rements.For example,a curriculum review in China determined that the level of requirements for the Information Science and Technology curriculum was above that of mathematics and science subjects,so the expectations had to be revised.The curriculum also had to cater for a wide range of contexts,and

262、the unique opportunities and challenges of both urban and rural settings.A pilot study of one part of the MIT DAILy Curriculum found that according to the teachers,the students seemed more engaged than usual,and what is and is not AI was the key to students understanding.Some teachers found the ethi

263、cs modules odd and confusing,but others embraced them.The use of hardware was perceived to be the most difficult component of the pilot to manage and required strong teacher investment and attention,particularly when the equipment did not work properly,although this helped students to learn importan

264、t skills like resiliency(Williams et al.,2021).Example:Qatar curriculum development foundations and principles16 Qatar Vision 203017 acknowledges technology as a key factor for a modern knowledge-based economy,and the Qatar National Curriculum Framework(QNCF)defines IT as a major school subject in g

265、rades 1-12 which aims to support childrens learning by offering beneficial opportunities in the areas of logic and mathematical thinking,language and communication,emergent literacy and creativity.To support these national policies,standards were created as a foundation of the national Computing and

266、 Information Technology curriculum through a collaboration between industry experts,the Qatar Ministry of Education and Higher Educations ICT experts team,and ICT supervisors at primary,preparatory and secondary schools.The standards were reviewed by the computer science experts and curriculum devel

267、opment experts of three Qatari higher education institutions.The curricula developed under the standards include a universal compulsory track for all grade levels and an elective high-tech track for senior high schools.Both tracks include AI learning outcomes related to algorithms,programming,ethics

268、 and social impact,and understanding and using AI tools and technologies.Students in the high-tech track also engage in developing AI technologies.The intention is to review the standards periodically to ensure that they capture current technologies and trends in computer science and IT.Ensuring the

269、 curriculum is not dependent on specific technologies,platforms or applications is another step toward guaranteeing the sustainability of the standards over time.The curriculum framework further suggests incorporating teacher feedback and international best practice reviews to make additional adjust

270、ments that bolster its effectiveness.The development of Qatars Computing and Information Technology curriculum standards was guided by five main principles:1.Alignment to a national curriculum framework,including the competencies,values,aims,principles and cross-cutting issues,with competencies 16 I

271、nformation in this section is sourced from the Qatar Computing and Information Technology Curriculum 2018 produced by the countrys Ministry of Education and Higher Education.It can be provided on request via email.17 See https:/www.gco.gov.qa/en/about-qatar/national-vision2030K-12 AI curricula A map

272、ping of government-endorsed AI curricula24explicitly linked to national standards.The curriculum covers knowledge,skills and attitudes with a focus on:Computer science principles and practices,namely programming,robotics and AI;Digital literacy,defined in the curriculum as the creative and productiv

273、e use and application of computer systems,including aspects of ethics,intellectual property and e-safety;The promotion of soft skills,in this case as defined by the American Association of School Librarians:collaboration,communication,teamwork,critical thinking,problem-solving and decision-making.2.

274、Spiral development,so that concepts reappear at different grade levels with increasing difficulty and greater depth with each iteration.At the same time,the skills development must be coherent and efficient,so that both needless repetition and academic gaps are eliminated.3.Student-centred learning

275、and hands-on project-based approaches.Computational thinking is a core element,and students are expected to leverage the process of abstraction,automation and analysis as a new approach to problem-solving,beginning with an understanding of the basic principles of computational thinking in grade 1.4.

276、Computing language,hardware and platform independence,meaning the curriculum does not rely on a particular provider,brand or programming language but seeks to cover a wide range of tools and technologies that students encounter in reallife.K-12 AI curricula A mapping of government-endorsed AI curric

277、ula25Curriculum integration and managementCurricula are integrated into existing education systems through a number of different models:y Discrete AI curricula are developed in an independent subject category within the national or local curriculum framework.These curricula have their own time alloc

278、ations,textbooks and resources,as in the case of Chinas Foundations of AI under Information Science and Technology for grades10to12.y Embedded AI curricula are developed and contained within other subject categories in the national or local curriculum framework.AI most commonly becomes a topic withi

279、n ICT or computer science but may alternatively be part of language,mathematics,science or engineering(see Figure 1).In the Republic of Korea,two elective AI subjects have been developed,one falling within the mathematics subject group and the other in technology and home economics.Curricula can als

280、o be designed to be embedded flexibly into any subject depending on teacher capacity and interest.This is the case for the MIT DAILy Curriculum.y Interdisciplinary AI curricula are implemented in systems with particular mandates for cross-subject work and associated time.These curricula target AI le

281、arning outcomes through project-based learning involving multiple subject areas.An example is seen in Portugals curriculum frameworks,which feature autonomous curriculum domains,or projects that must engage two or three disciplines in an interdisciplinary approach.In the UAE,AI is integrated into a

282、range of subjects including ICT,science,maths,language,social studies and moral education.y Multiple-modality AI curricula have core requirements which are implemented during school time and supported by traditional resources such as facilitator guides and textbooks,but also leverage informal learni

283、ng opportunities such as out-of-school resource networks and national or international competitions.An example of such a curriculum is the IBM-CBSE AI Curriculum for Grade XI&XII,which provides a gradual transition from guided to independent learning and links to competitions and industry mentorship

284、.y Flexible AI curricula can be implemented through one or more integration mechanisms at the discretion of regions,school networks or individual schools.Examples include the ATL AI modules curriculum in India,which can be embedded,interdisciplinary or delivered through out-of-school models such as

285、extracurricular activities;and Digital Skills in Saudi Arabia,which can be implemented as either a discrete or embedded curriculum.For some curricula,the embedding mechanisms are at the discretion of regions,schools or networks.These include the French-speaking Belgiums IT Repository(2nd and 3rd deg

286、ree technical transition),and Germanys Algorithmen erkennen und formulieren Identifying and Formulating Algorithms curriculum.Curricula can also be compulsory,meaning all students must participate;or elective,meaning that students choose to participate.In some curricula,such as Chinas Information Sc

287、ience and Technology,certain modules are compulsory and others are elective.Figure 1.Number of AI curricula by integration type(n=27,multiple responses possible)Number20Required subjectElective subjectPart of a required ICT/IT subjectPart of the elective ICT/IT subjectInterdisciplinary or

288、 cross-curricular subject Extra-curricular activity Source:UNESCO(2021b)K-12 AI curricula A mapping of government-endorsed AI curricula26One important point raised was that AI curricula,and ICT curricula more broadly,should not be heavily reliant on one particular technology,as it is important to di

289、versify the skills being developed across different platforms and providers.Some countries such as Austria and China emphasize an agnostic approach to technology,which means that the curriculum is not linked to any particular brands,devices or programming languages.These curricula therefore aim to e

290、nsure(i)that teacher training is firmly rooted in theory,ensuring an understanding of underlying principles that can be applied across a range of technologies;and(ii)if particular hardware or software is used,teachers and learners are introduced to multiple choices and be engaged with different prov

291、iders of AI tools.18 Boxplots show the distribution of data including the minimum value,first quartile value,median value,third quartile value and maximum value.The mean value is displayed as an X,and outliers appear as dots above the boxplot.Allocation of curriculum hoursSurvey respondents were ask

292、ed to provide the total number of learning hours for each of four grade levels:early primary,covering grades K-2;late primary,covering grade 3 to the end of primary school;middle school,covering grades 7 to 9 for most countries;and senior or high school,covering grades 10 to 12 for mostcountries.The

293、 total time allocations for curricula ranged from two to 924 hours,spread over between one and twelve grade levels.A boxplot18 of the time commitment per grade(see Figure 2)shows that the allocations varywidely.Figure 2.Time allocation per year of AI curricula,n=22 2218058Hours%5010015020

294、0250Source:UNESCO(2021b)Two outliers,the Computing and Information Technology(High Tech Track)curriculum in Qatar and the IT Repository curriculum in Belgium average more than 200 hours per year.The average of 58 hours is more than twice as high as the median of 21 hours,revealing that there is a cl

295、uster of curricula which require relatively few hours of engagement with AI.In fact,five of the 22 curricula which provided time allocations require fewer than 5 hours of AI study per year,while five require 150 hours or more.Those demanding 150+were either industry-developed curricula(two of the fi

296、ve)or high-technology elective tracks(also two of the five).Those requiring few hours of AI study were all embedded within other subjects.Curricula were most likely to target senior or high school learners,and the proportion of curricula engaging each educational level increases steadily from early

297、primary to the upper grades(see Figure 3).K-12 AI curricula A mapping of government-endorsed AI curricula27Figure 3.Per cent of curricula engaging each grade level(n=27)Early Primary Late Primary Middle/Junior School Senior/High SchoolPercent50403020100Source:UNESCO(2021b)The total hour c

298、ommitment for curricula per grade level ranged from 1 to 680 hours.In grades K-2,AI was most likely to be integrated into other subjects and have no specific time allocation.Only Qatars Computing and Information Technology dedicated a specific time allocation to K-2,100 total hours.Across grades 3 t

299、o 6,an average of 156 hours was reported.The average time commitment as a whole for middle school(grades79)was 109 hours,and for high schools(grades 1012)the average was 153.5 hours.The average hours per grade were relatively static in K9:33.3 hours in grades K2;39in grades 36;and 36.3 in grades 79.

300、In high schools,the average time commitment per grade increases to 51.2 hours.Essential conditions for supporting AIcurriculaSurvey respondents were asked how essential conditions were planned and prepared to support the design and implementation of the AI curriculum.The seven options that were pres

301、ented in the questionnaire are outlined in Table 7.Multiple responses were possible,and a free-response option was also provided.Table 7.Essential conditions for supporting AI curriculaResponse optionsCommentResearch or a needs analysisReferred only to research or a needs analysis related to the imp

302、lementation of the curriculum.Development of resources for teachersTextbooks and lesson plans were given as examples.Teacher training Respondents were asked about training specific to the AI curriculum and the resources facilitating it.Hiring of additional staff/capacityReferred to the recruitment o

303、f more paid teachers to implement the curriculum.Engagement of the private or third sectorRather or in addition to extra school staff,some countries engaged private or third-sector organizations as part-time trainers in or for schools.Infrastructure upgrades at schoolsReferred to the provision of ha

304、rdware and/or internet connections for schools in relation to the AI curriculum.This includes items such as computer labs and servers.Procurement of additional resources for schools or classroomsBuying in classroom kits,coding resources,AI tools,etc.Source:UNESCO(2021b)K-12 AI curricula A mapping of

305、 government-endorsed AI curricula28The responses show that implementing an AI curriculum requires a range of adjustments to the resources and human capacity of education systems(see Figure4).A majority of curricula were supported through the development of teachers resources and training(89per cent)

306、;15(56per cent)through preliminary research or a needs analysis;13(48per cent)through investment in school infrastructure upgrades;19 Sourced from:CBSE Artificial Intelligence(Ministry of Education,India,2020);Artificial Intelligence Curriculum,Class 9 Facilitator Handbook(CBSE and Intel,2019);and i

307、nterviews with and presentations from representatives of IBM,Intel,the 1M1B Foundation and Microsoft.Note that this information may not represent the official views of the Government of India.20 For example,1M1B supports the implementation of the AI Youth Skills curriculum in partnership with CBSE a

308、nd IBM.See https:/ cent)through engagement of the private or third sector;and 11(41per cent)through procuring additional resources for classrooms.The area emphasized the least was hiring additional school staff to implement the curriculum,but this was still a notable activity reported for 8 of the A

309、I curricula included in this study(30per cent).Figure 4.Support for implementation undertaken020406080100Research or Needs Analysis Develop teacher resources Teacher training Engage private/third sector teachers or trainers School infrastructure upgradesProcure additional classroom resourcesHire add

310、itional school staffSource:UNESCO(2021b)Example:The introduction of AI by the CBSE in India19In 2019,Indias Central Board of Secondary Education(CBSE)announced AI as an optional subject in the more than 22,000 schools under its jurisdiction,with a goal of ensuring that future citizens of India under

311、stand AI and are able to deploy it to address local and global problems.The curriculum focuses on learning through doing and provides opportunities for students to learn AI by using it to build solutions to community challenges(CBSE,2020).In support of the curriculum,the CBSE partnered with industry

312、 providers including IBM,Intel and Microsoft to develop training and support materials and content.NGOs also support the delivery of the curriculum.20 To prepare for implementation,teacher and mentor training,and materials including facilitator guides,multi-disciplinary lesson plans and textbooks we

313、re created for grades 812.The CBSE also pursued a number of events with the overall aim of smoothing the integration of AI into schools.These included competitions,virtual symposiums to give youth opportunities to explore AI technologies,and three-day AI-thon camps where students execute a project d

314、esign and prototype cycle using AI to solve an identified community challenge.More than 10,000teachers and 120,000 students have been trained in AI through various partnership activities likethese.The AI curriculum is integrated as an elective or interdisciplinary subject in self-selected schools.Th

315、e CBSE circulates an invitation to all schools to participate in this curriculum,and school administrators submit an application to the CBSE in response to the opportunity.Schools then select teachers for training,plan for AI to be included in the school timetable,and strategize around the interdisc

316、iplinary integration of AI based on common themes,such as improvement in food resources,the example in Figure 5(CBSE and Intel,2019).K-12 AI curricula A mapping of government-endorsed AI curricula29Figure 5.Thematic approach to the interdisciplinary integration of AI into the curriculumAI Integratio

317、n using Google Story SpeakerAI Integration using Computer VisionEnglishWrite a newspaper article suggesting strategies to improve rood production in the country.A discussion-With the population rise in India,more farmland is needed,while India is already intensively cultivated.Do you think Artificia

318、l Intelligence is a way to solve this problem?ScienceHigher yields of foodWhat do we do to get higher yields in our farms?Case study -Why can we not make do with the current levels of agriculture production?MathematicsProblem solving The population of India is more than one billion people and we nee

319、d a quarter of a billion tonnes of food every year.What data will you collect to present a research report on these topics?GeographyDoes climate impact gain production?What ways can you suggest to predict the climate and protect crops?What ideas do you suggest to improve the natural irrigation syste

320、ms?AI Integration using Natural Language ProcessingAI Integration using Data ExplorationTheme class 9ScienceChapter 15Improvement in Food ResourcesSource:CBSE and Intel(2019)At the school level,teacher training is facilitated by industry and/or implementation partners using bespoke courses and mater

321、ials.Various providers such as IBM and Microsoft have developed textbooks in line with the CBSE curriculum aims.The curriculum integration,syllabus,pedagogy,approach,and procurement of necessary resources are also managed at the institutional level.Moreover,schools are expected to engage stakeholder

322、s,especially students and parents,to ensure they understand the rationale and aims behind the integration of AI into the curriculum.Figure 6 outlines the relationship between the CBSE,schools and stakeholders.Figure 6.AI implementation actors and proceduresACBSELevelStakeholderLevelSchoolLevelSchool

323、LevelAIAI Sensitization For PrincipalsCurriculum IntegrationSyllabusPedagogy and Approach Arranging ResourcesPlanning Modalities forSchool ImplementationTraining of Teachers in AI In collaborationRationale of AI for Learning BeyondSource:CBSE and Intel(2019)K-12 AI curricula A mapping of government-

324、endorsed AI curricula30AI curriculum contentMain categories of AI curriculum contentThis analysis covers nine topic areas of AI curricula:algorithms and programming;data literacy;contextual problem-solving;the ethics of AI;the societal implications of AI;applications of AI to other domains;understan

325、ding and using AI techniques;understanding and using AI technologies;and developing AI.As outlined in Table 8,these nine topic areas are grouped into three categories:AI foundations;ethics and social impact;and understanding,using and developing AI.The survey respondents were asked to provide inform

326、ation on the time and percentage allocation for these topic areas,which are shown inTable 9.Table 8.AI curriculum areasCategoryTopic areaCompetency and curriculum considerationsAI foundationsAlgorithms and programmingTogether with data literacy,algorithms and programming can be viewed as the basis o

327、f technical engagement with AI.Data literacyA majority of AI applications run on big data.Managing the data cycle from collection to cleaning,labelling,analysis and reporting forms one of the foundations for technical engagement with using and/or developing AI.An understanding of data and its functi

328、ons can also help students understand the causes of some of the ethical and logistical challenges with AI and its role in society.Contextual problem-solvingAI is often framed as a potential solution to business-related or social challenges.Engaging at this level requires a framework for problem-solv

329、ing in context,encompassing things like design thinking and project-based learning.Ethics and social impactThe ethics of AI Regardless of technical expertise,students in future societies will engage with AI in their personal and professional lives many do so from a young age already.It will be impor

330、tant for every citizen to understand the ethical challenges of AI;what is meant by ethical AI;concepts such as transparent,auditable,and fair use of AI;and the avenues for redress in case of unethical or illegal use of AI,e.g.that which contains harmful bias or violates privacy rights.The social or

331、societal implications of AIThe social impacts of AI range from requiring adjustments to legal frameworks for liability,to inspiring transformations of the workforce.Survey respondents were asked about the extent to which their curricula targeted these issues.Trends such as workforce displacement,cha

332、nges to legal frameworks,and the creation of new governance mechanisms were given as examples.Applications of AI to domains other than ICTAI has a wide range of applications outside of computer science.The survey asked participants whether and to what extent AI applications in other domains were con

333、sidered.Art,music,social studies,science and health were given as examples.Understanding,using and developing AIUnderstanding and using AI techniquesThis area included(1)the extent to which theoretical understandings of AI processes were developed(e.g.defining or demonstrating patterns,or labelling parts of a machine learning model);and(2)the extent to which students used existing AI algorithms to