《北京师范大学智慧学习研究院:2022年全球中小学人工智能教育支撑环境白皮书(英文版)(34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年全球中小学人工智能教育支撑环境白皮书(英文版)(34页).pdf(34页珍藏版)》请在三个皮匠报告上搜索。
1、International Report ofSupportive Environment on ArtificialIntelligence Courses in K-12 Smart Learning Institute of Beijing Normal University(SLIBNU),2022Supported by Alibaba Cloud Computing Ltd.Rights and PermissionsInternational Report of Supportive Environment on Artificial Intelligence Courses i
2、n K-12This publication is available in Open Access under the Attribution-ShareAlike 4.0(CC-BY-SA 4.0)license (http:/creative commons.org/licenses/by-sa/4.0/).Please cite the work as follows:Huang,R.,Liu,D.,Own,C.M.,Zhang,J.B.,Li,Y.Y.,Chen,L.J,Hu,X.,Fan,L.,H,G.J.,Yao,Y.J.,Dai,Z.L.,Wang,J.X.,Chen,H.Y.
3、,Wei,Y.A.,Zhang,X.Y.,(2022).International Report of Supportive Environment on Artificial Intelligence Courses in K-12.Beijing:Smart Learning Institute of Beijing Normal University.2022,Version 1.01With the development of big data,computing power,and machine learning algorithms,the application of AI
4、technology is becoming more and more common in various industries and peoples daily life.The integration of AI and education is becoming closer and closer.How to use AI technology to improve education quality and teaching efficiency,and how to cultivate professionals who master AI technology have be
5、come the focus of global attention.In 2019,the president of the Peoples Republic of China Xi Jinping to International Artificial Intelligence and Education Conference pointed out that grasp the global AI development situation,find the breakthrough and main direction,cultivate a large number of innov
6、ation ability and cooperation spirit of AI high-end talents,is the important mission of education.As an important result of the meeting,more than 100 member countries and UN institutions,AI,civil society and the private sector of about 500 representatives jointly published the BEIJING CONSENSUS AI a
7、nd education(Beijing consensus),The consensus is committed to implementing appropriate policy strategies through AI and education system,comprehensive innovation education,teaching,and learning methods.In addition,the consensus is to use AI to speed up the construction of an open and flexible educat
8、ion system to ensure that people enjoy fair,suitable for everyone,and high-quality lifelong learning opportunities.This will thus promote the realization of the Sustainable Development Goals and a community with a shared future for mankind.At the opening ceremony of the AI and Education Internationa
9、l Forum,Chinese Education Minister Huai Jinpeng stressed the importance of AI in education,and mentioned that AI will give education power,change education and innovate education,which will undoubtedly create a better future for everyone else.To promote youth AI education and analysis of the support
10、ive environment,the National Engineering Research Center for Cyberlearning and Intelligent Technology joined the Smart Learning Insitute of Beijing Normal University in carrying out the International Report of Supportive Environment on Artificial Intelligence Courses in K-12.This report aims to,thro
11、ugh all kinds of data collection,summary,analysis,and comparison,for academia,education,and industry to provide a comprehensive understanding of the supportive environment for AI courses in K-12.The report also discusses the key elements that affect K-12 learning in AI courses.For instance,it includ
12、es teaching techniques for K-12 students,such as assigning independent learning,analyzing possible trends for the future AI development of youth education,and providing sufficient teaching resources for teachers,scholars,or other personnel.This report contains six chapters,briefly described as follo
13、ws:The first chapter is an overview based on clarifying the concept of AI and reviewing the development situation of AI in education.This chapter expounds on the two relationships between AI and education:AI as an auxiliary education tool and AI as learning content.The Beijing Consensus summarizes t
14、he impact of AI on education development and puts forward five critical elements of the supportive environment for AI courses in K-12.The second to fifth chapters introduce the elements of the supportive environment in K-12 AI courses.The second chapter introduces open educational resources,the teac
15、hers development online community,analyzes the benefits of using open education resources,introduces some AI learning platforms and online communities for educators.The third chapter summarizes the current situation of K-12 AI textbooks,presents the selection of AI textbooks for different ages,and i
16、ntroduces some AI teaching practices.The fourth chapter discusses the challenges and prospects of the K-12 AI algorithm.It also summarizes the datasets that may be involved in algorithm teaching.The fifth chapter presents the classification and characteristics of AI teaching laboratories and lists s
17、ome examples.Finally,chapter 6 is the conclusion and recommendation of the report,which summarizes the overall situation of the supportive environment of AI education.Due to the vital influence of AI technology,countries invest a lot of resources in auxiliary and AI education.The application of AI i
18、n education and its education resources will be more and more.The case listed in the white paper is only part of the case collected so far and can lay a foundation for readers to understand the basic situation in the field.I believe this field will continue to grow and have a better future developme
19、nt.Huang RonghuaiProfessor of Beijing Normal UniversityDirector of National Engineering Laboratory for Cyberlearning and Intelligent TechnologyPrefacePreface-International Report of Supportive Environment on AI Courses in K-122International Report of Supportive Environment on AI Courses in K-12-Tabl
20、e of contentsInternational Report ofSupportive Environment on ArtificialIntelligence Courses in K-12Table of contentsPrefaceAn Overview of Artificial Intelligence and Education1.1 Definition and classification of artificial intelligence1.2 The current development of global artificial intelligence in
21、 education 1.3 AI as a tool and learning content 1.4 The impact of artificial intelligence on education development 1.5 Key factors in the implementation of AI education 1.Open Educational Resources and Teacher Development Online Community2.1 Introduction of open educational resources2.2 Open educat
22、ional resources for AI learning2.3 Benefits of using open educational resources 2.4 Teacher Development Online Community 2.K-12 AI Teaching Materials and Teaching Methodology3.1 Current status of K-12 AI textbooks 3.2 K-12 textbook selection3.3 Teaching practice of artificial intelligence 3.K-12 AI
23、Algorithm and Dataset4.1 An AI algorithm suitable for K-124.2 Challenges faced in the teaching of K-12 AI algorithm4.3 The Outlook on Algorithm Teaching in K-12 AI Education 4.4 Dataset suitable for K-12 AI education 4.AI Laboratory for K-125.1 Classification and characteristics of AI laboratories 5
24、.2 K-12 AI lab example5.Conclusion and Recommendation6.1 The Conclusion on K-12 AI Education 6.2 K-12 AI Education Recommendation 6.References7.293344578999An Overview of Artificial Intelligence and Education-International Report of Supportive Environment on AI Cours
25、es in K-121.An Overview of Artificial Intelligence and EducationArtificial intelligence(AI)is changing every aspect of our lives.Under the influence of AI,transportation,healthcare,logistics,finance,and industrial manufacturing,will develop more productive and cost advantages and,more importantly,pr
26、ovide better services,where education will be one of the most affected areas1.The use of AI in According to a Global Market Insight report,AI use in education is expected to explode to$6 billion in the next six years.It turns out that education systems and institutions need a lot of reform,and AI wi
27、ll take a significant role in the developing process.1.1 Definition and classification of artificial intelligence1.2 The current development of global artificial intelligence in educationArtificial Intelligent(abbreviated as AI)is a generic term for a machine or computer program capable of performin
28、g tasks or activities(such as planning,problem-solving,identifying patterns,and logical actions)that require features of human intelligence2.While the term was first coined in the 1950s,AI has exploded,driven by the Internet,big data and cloud storage,and more powerful computing and algorithms.In th
29、e field of AI research,the application scope of AI is divided At present,the governments of most of the worlds developed economies have determined the potential of AI as a competitive element of its future labor force and economy,all promoting the popularization of relevant education in different wa
30、ys.In 2017,Information Technology Curriculum Standard for Chinese Ordinary HighSchools was released,and AI became included in the curriculum system of Chinese high school students.In May 2019,Xi Jinping,the president of the Peoples Republic of China,sent a congratulatory letter to the International
31、Conference on Artificial Intelligence and Education.He pointed out that it is an important mission of education to grasp the global development trend of AI,identify the right breakthrough and main direction,and cultivate a large number of high-end AI talents with the innovative ability and cooperati
32、ve spirit.In December 2021,Huai Jinpeng,Chinese Minister of Education,proposed to fully integrate AI education into education at all levels and of all types to improve students digital skills and digital literacy.In April 2022,Information Technology Curriculum Standards for Chinese Compulsory Educat
33、ion were released,and AI added to the learning content of grades 7-9 of Compulsory education in China.In May 2018,the American Association for Association for the Advancement of Artificial Intelligence(AAAI)and the Association of Computer Science Teachers into three stages3:Artificial Narrow Intelli
34、gence(ANI),Artificial General Intelligence(AGI),and Artificial Super Intelligence(ASI).Artificial Narrow Intelligence can only complete specific tasks,not compound tasks.Artificial General Intelligence refers to the AI that can perform all human work.Artificial Super intelligence is a high degree of
35、 AI autonomy,far more than human beings.(Computer Science Teachers Association,CSTA)jointly formed a working group to launch the US K-12 AI Education Action4.Since 2016,Japan has been exploring and developing AI and programming education,proposing to help all children build the AI thinking needed in
36、 this era.In 2014,the Ministry of Education launched a new computer syllabus that teaches children to write simple computer programs,store and retrieve data from the age of 5,learn computer programming languages and solve computer failures from the ages of 11 to 14.Countries,including the United Sta
37、tes,Japan,and the UK,are now beginning to explore and practice the combination of AI with K-12 education.In February 2022,UNESCO released the K-12 Artificial Intelligence Curriculum.Officially Endorsed AI Curriculum Design Guide,which describes the scope and length of the nine categories of AI cours
38、es covered by existing K-12 AI curricula,as well as the expected learning outcomes for each category,summarizes the prerequisites needed to implement the curriculum,including teacher training,learning tools,and environments,and provides pedagogical recommendations5.The combination of AI and K-12 edu
39、cation is the trend of the Time,but how to do it is a huge challenge for educators and policy-makers who have to make todays students thrive in a very uncertain tomorrow.41.3 AI as a tool and learning content1.4 The impact of artificial intelligence on education developmentIn the field of education,
40、there are two relationships between AI and education6.One is the integration of AI and education,namely Artificial Intelligence in Education(AIED).The second is to cultivate AI talents that take AI as educational content.AIED has two research objectives:one is to make a comprehensive and in-depth ap
41、plication of AI technology in education to promote education reform and development.The other is to use AI technology more systematic and micro,In May 2019,The Beijing Consensus on Artificial Intelligence and Education10 was released by UNESCO,leading the 2030 education agenda Sustainable Developmen
42、t Goals 4,committed to implementing appropriate policy strategies through AI and education system.This includes comprehensive innovation education,teaching and learning methods,and the use of AI to speed up the construction of open and flexible education systems.The ultimate goal is to ensure fair a
43、nd suitable lifelong learning opportunities for everyone and to develop goals that can be shared with future humankind.This report and the guidelines proposed by UNESCO have and reveal the principles and mechanisms of learning,to create conditions for learners to master certain knowledge7 effectivel
44、y.With the research and exploration of AIED,the application of AI in education is also becoming more abundant,and the main application types are shown in Table 1.a wider audience and are closer to front-line teachers.AI has several effects on education,which are mentioned in the Beijing Consensus.Th
45、is report extends a more detailed content according to the Beijing Consensus.Here are several effects of AI on education as shown in Table 2.As a field that all developed economies attach great importance to,AI has gotten the attention on the implementation of AI education in primary education to cu
46、ltivate people those who have innovative talents in an intelligent society.The basic education stage in learning the content of AI is to let students understand the development and concept of AI and learn the implementation of typical AI algorithms.This learning process includes building a simple AI
47、 application module,experience design,and realizing the basic process and method of a simple intelligent system,enhancing intelligent technology service for developing human responsibility9.International Report of Supportive Environment on AI Courses in K-12-An Overview of Artificial Intelligence an
48、d EducationAPPLICATION TYPESPECIFIC SCENARIOS,APPLICATION TECHNOLOGY Intelligent and adaptive learningVirtual teachers created by integrating intelligent adaptive learning technology can penetrate the whole teaching process and support personalized teaching.Each student can learn at their own pace,i
49、ncreasing learning efficacy and motivation.Man-machine interactionSmart Energy processing and search technologyDual-teacher classImage recognitionSpeech evaluationIntelligent language processing and speech recognitionIntelligent language processing applicationsBased on language processing,some synta
50、x framework can be establishedPhoto-based question searchComputer vision and image recognitionTABLE 1:TYPES OF AI APPLICATIONS IN EDUCATION5An Overview of Artificial Intelligence and Education-International Report of Supportive Environment on AI Courses in K-12INFLUENCEACTION CONTENT Improve educati
51、on management and education governance capacityBe cognizant of the breakthrough in the use of data in transforming evidence-based policy planning processes.Consider integrating or developing AI technologies and tools that are relevant for upgrading education management information systems(EMIS)in or
52、der to enhance data collection and processing,making education management and provision more equitable,inclusive,open and personalized.From Article 10 of the Beijing ConsensusHelp teachers to empower teachingBe mindful that while AI provides opportunities to support teachers in their educational and
53、 pedagogical responsibilities,human interaction and collaboration between teachers and learners must remain at the core of education.Be aware that teachers cannot be displaced by machines,and ensure that their rights and working conditions are protected.From Article 12 of the Beijing ConsensusSuppor
54、t the development of a lifelong learning system for allReaffirm that the guiding principle for achieving SDG 4 is lifelong learning,which encompasses formal,non-formal,and informal learning.Adopt AI platforms and data-based learning analytics as key technologies in building integrated lifelong learn
55、ing systems to enable personalized learning anytime,anywhere,and potentially for anyone,with respect for learners agency.Exploit the potential of AI to enable flexible learning pathways and the accumulation,recognition,certification,and transfer of individual learning outcomes.From Article 20 of the
56、 Beijing ConsensusBe mindful of the need to give appropriate policy attention to the needs of older people,especially older women,and to engage them in developing the values and skills needed for living with AI in order to break the barriers to digital life.Plan and implement well-funded programs to
57、 equip older workers with skills and options that enable them to remain economically active for as long as they choose and to engage in their societies.From Article 21 of the Beijing ConsensusRealize personalized learning servicesConsider also introducing new models for delivering education and trai
58、ning in different learning institutions and settings that can be enabled by the use of AI,in order to serve different actors such as students,teaching staff,parents,and communities.From Article 11 of the Beijing ConsensusHelp to improve skills to precisely match employment needsBe mindful of the sys
59、temic and long-term transformation of the labor market,including its gender dynamics,due to AI adoption.Update and develop mechanisms and tools to anticipate and identify current and future skill needs in relation to AI development,in order to ensure the relevance of curricula to changing economies,
60、labor markets,and societies.Integrate AI-related skills into the school curricula and qualifications of technical and vocational education and training(TVET)and higher education,taking into consideration the ethical aspects and interrelated humanistic disciplines From Article 17 of the Beijing Conse
61、nsusTABLE 2:THE IMPACT OF ARTIFICIAL INTELLIGENCE ON EDUCATIONAL DEVELOPMENT1.5 Key factors in the implementation of AI educationAI used in K-12 education can be divided into two parts.One is to use AI technology,such as intelligent perception,intelligent education evaluation,intelligent recommendat
62、ion of digital education resources,and realize personalized learning,that is,the application of AI technology in various education scenarios11.The other is to carry out AI education,namely primary and middle school students accept artificial intelligence-related content learning,AI theory,computing
63、thinking,programming,and other related knowledge and skills1.The former AI technology can be seen as a help to K-12 education,and the second is to specifically describe learning about AI technology.Therefore,based on these two relationships,the five key elements of the primary and secondary school c
64、ourses are extended,which are the necessary factors to develop K-12 AI education.As shown in Figure 1:Figure 1:Key elements of AI teaching(v)AI laboratory(iv)Open-source algorithms,tools,and datasets(iii)Teacher Development online community(ii)OER for AI(i)AI teaching materialsKey elements6(i)AI tea
65、ching materials:it refers to the textbook used for AI teaching.It is compiled according to the curriculum standards and systematically reflects the content of the subject.The textbook is the embodiment of the curriculum standards,which is different from the general books,usually divided by school ye
66、ar or semester,divided into units or chapters13.(ii)OER for AI:refers to all kinds of open educational resources enhanced by AI technology,using open authorized teaching materials,usually digital materials,which can be used for teaching,learning,and research14.(iii)Teacher Development Online Communi
67、ty:It is the online community for teacher to learn for his job.A series of activities enable teachers to reflect on practice with colleagues,share expertise in a distributed knowledge framework,and establish a common understanding of new teaching methods,standards,and curriculum15.(iv)Open-source al
68、gorithms,tools,and datasets:Algorithms are a series of clear instructions to solve problems,and algorithms represent the systematic method to describe the problem-solving policy mechanism16.Tools refer to AI development tools,such as TensorFlow,flying paddle(Paddle Paddle),and other AI open-source f
69、ramework.Datasets refer to the sample data used for the training of the AI algorithm,such as MNIST for character recognition,ImageNet for image recognition processing,etc.(v)AI Laboratory:Laboratory for K-12 learning of AI content,including equipment for personal workstations and other equipment for
70、 instructional laboratories17.Among them,element one and element two are necessary for the learning content in the relationship between artificial intelligence and education,while elements three,four and five cover the necessary elements of teaching tools.The subsequent sections of this report will
71、outline the global development status and issues of these five key elements.International Report of Supportive Environment on AI Courses in K-12-An Overview of Artificial Intelligence and Education72.Open Educational Resources and Teacher Development Online CommunityThis chapter provides an overview
72、 of the Global K-12 AI Open Education Resource(Open educational resources,abbreviated as OER)and the Teacher Development OnlineCommunity.The OER is any type of educational material in the public domain18.More and more schools around the world begin to support the digital education model.Although the
73、 use of open educational resources cannot allow learners to get degrees,it is of great help to open education and distance teaching,and also helps to make the distribution of educational resources more evenly,indirectly making up for the gap in educational resources between countries19.At present,th
74、e teachers development online community has become an essential platform for teachers professional development worldwide.AI professional education community began to appear in the international world;it enables teachers to collaborate in learning,share experiences,and help solve problems.More import
75、antly,the community helps to cultivate internationally competitive and talented teachers and provides more means to increase youth information literacy.2.1 Introduction of open educational resourcesThe term OER were initially proposed at a conference hosted by UNESCO in 2000 and were promoted in the
76、 context of providing free access to educational resources worldwide.OER use publicly authorized textbooks,usually digital materials,which can be used for teaching,learning,and research20.They are released as open licenses(i.e.,knowledge sharing),describing how to use,reuse,adapt,share,and modify th
77、e material for specific needs.Those contents include textbooks,lecture notes,syllabuses,assignments,and exams.OER are a critical forefront of expanding access to quality education,reducing the chance of getting overly complicated teaching materials,and encouraging teachers,students,and educational i
78、nstitutions to participate in creation and collaboration actively.Educational institutions and organizations,including some of the worlds most prestigious universities and colleges,are working to create and disseminate this resource to support academics and students.MIT Open Course Ware has played a
79、 considerable role in promoting the development of global open educational resources.With over 2,500 online publications on MIT course materials,the site freely shares knowledge with learners and educators around the world,including information technology types of courses and lectures.The British Op
80、en University also significantly impacts the construction of open education resources in the field of distance education,which can support more than 170,000 students.British Open universities break the constraints of time,space,teaching environment,and economic conditions through distance education
81、and open educational resources to promote teaching equity.OER Initiative provides open learning opportunities for high-quality education resources worldwide.From an institutional base of large or institutional-supported initiatives to many small activities,the number of programs,and programs related
82、 to OER has increased rapidly over the past few years.Open Educational Resources and Teacher Development Online Community-International Report of Supportive Environment on AI Courses in K-128The OLI offers dozens of university-level courses,as well as a platform to research and experiment with any a
83、spect of the learning experience.Dedicated to maintaining low cost courses in class and providing free courses for independent learners.The Connexions program provides educational content to millions of users each month to improve learning outcomes.There are thousands of textbooks that are easily ac
84、cessible online and can be downloaded too.Almost any device anytime,anywhere.Connexions2.2 Open educational resources for AI learningThere are few platforms suitable for K-12 to learn AI.The examples collected in this report are all open,content-rich,and interesting.More importantly,K-12 can fully u
85、nderstand the technology of AI and learn such AI thinking.As Dale Lane on the Machine Learning for kids platform once said in a speech:Its good to introduce machine learning to teenagers now,not just to people like developers or technicians.Some of them may continue to be the inventors of the next g
86、eneration of technology.They will invent the next machine learning OpenLearn currently has about 35,000 pages of high-quality resources,including more than 1,000 courses,with a simple mission:to reach all people who need free higher education.MIT OCW is a free open publication of material from thous
87、ands of MIT courses,covering the entire MIT curriculum.The MERLOT is an international community of educators,learners,and researchers.Multimedia Education resources for learning and online teaching,serving more than 900,000 students and over 47,000 teachers.system that can solve larger problems that
88、 we cant approach today.We need them to keep the technology forward,maybe introduce machine learning to them today to play,experience,understand,maybe it will inspire their imagination and start a lifetime of invention and creation.Here are a few examples listed as follows.OpenLearMIT OCWOLIMERLOTNA
89、MEACTION CONTENT URLMachine Learning for kidsIt is a free online learning platform for teenagers to train computers on how to accomplish the tasks given.It guides them step by step to create machine learning models,and the types that can be identified are text,numbers,images,and sound.The user first
90、 collects the examples you want to identify,then trains the computer to identify them,and finally makes a game or app that enables the computer to recognize them.https:/machinelearningforkids.co.uk/Code.orgAimed to enable every child in the world to receive computer science education.The group targe
91、ts the K-12 age group and customized free courses with content such as AI,which can be used to allow children aged 4-6 and 8-12 to start learning.Each course includes video tutorials and fun hands-on exercises to help children apply what they learn while maintaining their interest in learning.https:
92、/code.org/Khan AcademyThe computer programming program at Khan College is completely free,which leads children and students to learn computer-based,artificial intelligence-related courses.https:/www.khanacademy.org/Elements of AIThe program has spread the world with graduates from more over 170 coun
93、tries.Over 750,000 students were enrolled in the AI Elements course.About 40%of course participants were female,more than twice the average computer science course.The sites curriculum is divided into two categories,a simple introduction to AI and building AI applications.https:/ 3:K-12 PLATFORM FOR
94、 LEARNING AI International Report of Supportive Environment on AI Courses in K-12-Open Educational Resources and Teacher Development Online CommunityFigure 2:The Typical OER Project9NAMEINTRODUCE URLYouth Artificial Intelligence Innovation Initiatives(Yuanzhuo Plan)The aim is to explore the linkage
95、mechanism between the government,universities,and enterprises,integrate various forces and resources of the society,stimulate the interest of teenagers in using original AI algorithms to solve real and complex problems,serve the country,guide teachers,and cultivate students through public welfare.ht
96、tps:/ website supports and improves K-12 education and teaching,science and technology,and management.It also provides a free and practical guide for students to participate in AI creation for free.ISTE and GMs Classroom Practice AI Program Guide provide computer science teachers with innovative cur
97、riculum resources regarding AI in different grades and subject areas.https:/www.iste.org/areas-of-focus/AI-in-educationKQED TeachOffers a range of free-practice professional learning opportunities focused on digital media.Educators can develop skills in digital narrative,data visualization,and key m
98、edia use to support all curriculum areas,especially AI-related courses.These skills enable teachers to facilitate a learning environment,where students can create digital content,develop their communication and technical skills,and participate in encouraging deeper learning of critical thinking.Ther
99、e are analysis and evaluation classes,teacher video production,podcasts,and audio production for teachers to learn.https:/teach.kqed.org/AI for TeachersIt is a website dedicated to supporting the integration of AI knowledge throughout K-12 learning,providing educators with quality planning resources
100、 to address the challenges of integrating AI into existing courses.https:/aiforteachers.org/TABLE 4:THE TEACHER ONLINE DEVELOPMENT WEBSITE2.3 Benefits of using open educational resources2.4 Teacher Development Online CommunityEnsuring the equality and availability of education is the fundamental val
101、ue of the open education movement,and addressing these issues becomes increasingly important as we move to the future21.Open education addresses many issues that complicate education,including:The professional development of teachers is seen as standard-guided and rooted in teacher work.It focuses o
102、n students learning,and is suitable for the lifelong and situational form of teachers career development,the purpose is to develop,implement and share theknowledge between teachers,practice,and measure the online and offline activities of all students22.Outdated textbooks:Traditional educational mat
103、erials,such as textbooks,can only be as timely as their publication date.Unfortunately,the latest knowledge,especially in the AI technology area,will soon be obsolete.In the modern world,students and teachers who use outdated materials are in very disadvantaged positions.On the other hand,OER can be
104、 constantly updated,adjusted and improved,so that they are never obsolete.Cost:Setting up AI workstations and learning materials often have high prices,raising a painful problem for many students:having no materials or being heavily in debt to get them.Cost of education is a significant social probl
105、em.The goal of open education is to make high-quality,free online material,which is a direct answer to the question of cost.Access and Equality:Cost,geographical and structural inequalities in society often lead to educational opportunities and equity.Through open education,high-quality materials ar
106、e available at minimum(or no)costs,and with minimal access barriers,providing a level playing field for students to prepare for future life.Open Educational Resources and Teacher Development Online Community-International Report of Supportive Environment on AI Courses in K-12103.K-12 AI Textbook and
107、 Teaching MethodologyThis chapter explores the educational materials and teaching methodologies used in K-12,presenting a number of countries around the world on AI educational materials and extracurricular books,categorizing them by age group.Adolescents are the new generation living in an intellig
108、ent environment,their lives are closely related to AI,and children are eager to learn about AI.Therefore,it is important to popularize computational thinking and AI thinking methods among teenagers,and to cultivate their strong interest in AI research and problem solving using AI,in the hope that th
109、ey will be able to innovate technology in the future23.3.1 Current status of K-12 AI textbooksA comprehensive survey on various kinds of complicated books on AI at home and abroad has found that with the popularization of AI,several AI books suitable for teenagers have appeared.However,there are so
110、few countries that add AI courses to the basic education system as independent courses,there are so few AI textbooks applied to classroom teaching.In terms of content,kindergarten and primary school textbooks focus on practical life,and experience abstract concepts through cases to help students per
111、ceive AI;junior high school AI textbooks focus on programming teaching,programming as a link,experience,develop and understanding AI;high school textbooks go deep into programming and robot education,reflecting the idea of multidisciplinary integration24.At the level of teaching methods,each textboo
112、k emphasizes experience perception and the integration between disciplines.Because AI is a subject with strong manual practice,AI learning needs the corresponding practice environment.This has led to some textbooks relying on specific hardware and software platforms,and their universality needs to b
113、e improved.In addition,most of the AI textbooks are relatively difficult,and they lack of pilot data and effect evaluation.3.2 K-12 textbook selectionAI childrens books,including fun books,novels,activity manuals,and useful tutorials,guides,all help children prepare for their future life and develop
114、 their interest in AI technology25.This report believes that the AI reader suitable for K-12 students has five-article standards:The following are the results obtained after the research group screened the globally retrieved teaching books according to the above selection principles,as shown in Tabl
115、e 5.International Report of Supportive Environment on AI Courses in K-12-K-12 AI Textbook and Teaching MethodologyiiiiiiAdapt to the childrens age,development,and understandingCultivate children with good moralityivProvide useful information under the current trendsInspire AI thinking and creativity
116、vSuitable for repeated readingFigure 3:Selection Principles of AI Textbook11K-12 AI Textbook and Teaching Methodology-International Report of Supportive Environment on AI Courses in K-12TYPEAGE GROUP CHARACTERISTICSNAMEAUTHORPUBLISHERCONTENT INTRODUCTIONPopularize cognitionEarly children,primary sch
117、ool students age groupConcrete abstract concepts through cases,and help children perceive and understand AI.AI+MEReadyAI ReadyAI Independent Publishing(July 22,2020)The picture book for K-2 students respectively introduces the five concepts of AI perception,representation and reasoning,machine learn
118、ing,the interaction between people and AI,and social impact.The book is suitable for being a popular science book for childrens AI knowledge.Junior high school stageSpecial attention is paid to the cultivation of AI thinking,to experience,develop and understand AI.Artificial Intelligence:101 Things
119、you have to know today about the futureLasse Rouhiainen CreateSpace Independent Publishing Platform(January 29,2018)It covers many interesting and time-effective topics related to AI,including self-driving cars,robots,chatbots,and how AI affects the job market,corporate business processes,and conten
120、t across the industry.Suitable for serving as a junior high school popular science books.Senior school educationThe concept of AI has been thoroughly understood,and it focuses more on how to achieve the actual goals.Artificial Intelligence(High School Edition)Tsinghua UniversityTsinghua University P
121、ress(September 2020)In the book,the eight core basic modules of AI are fully sorted out,which can be used as high school AI curriculum textbooks,or as popular science books.The book is also equipped with exclusive website resources,providing programming help and advanced content to help people inter
122、ested in AI learn further.Thinking trainingEarly children,primary school stageIt is mainly the popularization of five concepts of AI,so that young children can also understand the concept of AI.How To Train Your RobotKen Goldberg et al(Supported by the NSF and the UC Berkeley Lawrence Science Instit
123、ute).Free to download(November 2019)The book is a childrens book for K-8,written by Bluma Goldberg in 2019,and is available for free download from online.The book introduces readers to cutting-edge robotics and AI in a highly accessible way and mimics real engineering practices such as repeated desi
124、gn,testing,and pass failure learning.Junior high school stageLet students understand the characteristics and scope of AI technology,the design process that engineers develop new devices,and how technology affects humans and societyArtificial Intelligence(Junior Edition)Fan Lei et alTsinghua Universi
125、ty Press(August 2020)The book gives a preliminary explanation of AI from the development and application of AI,data,algorithm,supervision and learning and other aspects.Junior high school students can understand the knowledge and technologies related to AI and stimulate their interest in AI.The book
126、 is suitable for junior high school students,and can be used as an AI entry textbook for junior high school students.TABLE 5:K-12 TEXTBOOK SELECTION12 International Report of Supportive Environment on AI Courses in K-12-K-12 AI Textbook and Teaching MethodologyTYPEAGE GROUP CHARACTERISTICSNAMEAUTHOR
127、PUBLISHERCONTENT INTRODUCTIONThinking trainingSenior school educationEmphasize the cultivation of ability and thinking,and understand,apply and innovate AI in a practical way.What to Think About Machines That ThinkJohn BrockmanHarper Perennial Press(October 2015)Nearly 200 short papers,written by re
128、nowned scientists,philosophers,and futurists,explore the question of human and machine intelligence and consciousness.The book is written for educated non-computer professionals,and a wide variety of views may cause high school students to think about other issues on AI,so it is suitable for high sc
129、hool students and above.Technical ability learningelementary school stageThrough simple practical cases,the abstract concept is clearly described,and then draw inferences from one example,to practice the knowledge just learned.AI future intelligent maker:artificial intelligence boutique course serie
130、s in primary and secondary schoolsWang Jiqing et alEast China Normal University Press(October 2018)This set of books allows students to perceive,recognize and apply the connotation and extension of AI,thus becoming interested in AI technology.Each book is aimed at a special topic,and each special to
131、pic contains 16-18 theme content and 2-4 activity courses.The knowledge of AI advances layer by layer,and students can form a systematic knowledge framework of AI.Junior high school stagePay attention to the cultivation of AI technology and engineering thinking,junior high school,the final thing is
132、the cultivation of thinking.How Smart Machines ThinkSean GerrishMIT Press(October 2018)Offering a fresh and modern vision of AI,machine learning,and deep learning about Jeopardy video game shows,Netflix,Starcraft,board games(e.g.,Go,chess,Sudoku),and self-driving cars.Great entry for engineers inter
133、ested in putting the AI in context.Senior school educationIn the critical period of cultivating AI technology ability,the high school textbooks are in-depth programming and robot education,which reflects the thought of multidisciplinary integration.Fundamentals of Artificial IntelligenceDr.Nisha Tal
134、agala et alAIClubThe book covers the basics of AI,and its mathematical and coding levels match the capabilities of middle and high school.Each chapter is equipped with exercises to help improve students practical operation ability.CONTINUED FROM TABLE 5:K-12 TEXTBOOK SELECTION13K-12 AI Textbook and
135、Teaching Methodology-International Report of Supportive Environment on AI Courses in K-123.3 Teaching practice of artificial intelligenceSupport school-wide pilot tests on the use of AI to facilitate innovation in teaching and learning,drawing lessons from successful cases and scaling up evidence-ba
136、sed practices.From Article 15 of the Beijing ConsensusIn the future,the deep integration of AI and education will force teachers to face and deal with enormous challenges.The role of the teacher will also undergo a significant transformation teaching knowledge to guiding the overall development of s
137、tudents moral,intellectual and physical development,and acting as a mentor and counselor6.Teachers under AI should also become the realizer of personalized education.Teachers should learn to use the student learning data provided by AI and analyze it accordingly to accurately obtain personalized dat
138、a of students in the learning process and implement personalized teaching and targeted teaching.Teachers information technology capabilities should also be accelerated,and through the construction of smart education systems and teaching platforms,teachers should realize accurate teaching,screen high
139、-quality teaching resources and teaching materials and upload them to the platform to provide optional teaching content and information for students.The following are specific examples of the AI teaching practice:(i)Project-based LearningProject-based learning(PBL)is a teaching methodology that rein
140、tegrates interconnected knowledge points into several relatively independent teaching programs,and organically combines theory and practice to mobilize learners with the help of teaching programs7.Project teaching comes from Dewey Do middle school(learn by doing)and Coopers experiential learning the
141、ory(experimental learning),etc,.The core idea is group cooperation to build physical orientation through real problems in the real world to capture students interest,guide students to analyze problems,find resources to design clear solutions in the process to stimulate students deep thinking and lea
142、rning,realize the understanding and internalization of knowledge.This methodology has changed the traditional teacher-centered education concept of teaching,emphasizing student-centered learning,and truly makes students become active builders of knowledge.The project points to students high-order th
143、inking ability,which mainly solves the problem of bad structure,so that students can perceive and understand knowledge in the process of project practice,and realize the improvement of problem analysis,logical reasoning,collaborative communication and other abilities.Case:Artificial intelligence rob
144、ot Cozmo programming thinking and interactive applicationIntroduction:The project is an AI experience summer camp designed by the founder of the Cozmo Robot Programming Language.Cozmo is an AI toy,the first smart toy robot to have visual,auditory,vocal,and programmable instructions and interaction w
145、ith people.The program is based on CMUs Calypso(Ocean Fairy,a rule-based programming language).Target Students:The high school students.Course content:Programmable instructions and interactive experiments for intelligent toy Cozmo.Students learn the reasoning framework of different programs,predict
146、the robots behavior to the program,and writes calibration programing when the program has mistakes;by writing Calypso programming language,making the robot complete specific actions and events,simulating real situations.Teaching Procedure:The course is about three weeks.Prior to the formal start of
147、the course,training on basic knowledge,including basic knowledge of Calypso programming language,project design thinking,and project planning training is provided.In the first week,students independently design the robot project theme,implementation plan and conception;the second week is R&D and exp
148、eriments,with practical programming training and simulation experiments around the project topics,independently designed by students.The theme of the third week is summary&refining.Based on the previous simulation experiment,summarize the project results,design the defense mode,and record the defens
149、e video.(ii)Problem-based learningProblem-based learning(PBL)is a contextual-based,student-centered teaching methodology that anchors learning into complex and interesting problems,usually group cooperation29.PBL will ask students to solve real-life problems,learn the scientific knowledge behind the
150、 problem,and form problem-solving skills.The role of teachers is to provide ways to obtain learning resources and appropriate guidance on learning methods.Unlike the traditional classrooms where teachers use a problem to lead students to the lesson topic and discard it right after or after completin
151、g the lesson and use problems to test students,in PBL,all learning activities build upon the initial problem.Theacquisition of personal knowledge is implicit in the process of problem-solving.Subject knowledge is the assistant serving problem-solving.Therefore,as the starting point of learning,the p
152、roblem is whether problem-based learning can be effectively implemented.14Case:Qingdao Artificial Intelligence Class30Introduction:The course is an open AI class in Laoshan District,Qingdao,China,in 2020.The speaker is Guo Chunlei,an information technology teacher from Laoshan District No.2 Experime
153、ntal Primary School in Qingdao.Target Students:grade 4 students.Course content:Basic principles of AI machine recognition facial recognition.Teaching Procedure:Students,can you help me quickly find all the photos of Liu Yankai from these images within 1 minute?This is the question the lecturer asked
154、 to the students in the beginning.The whole classroom is divided into three main parts.The first part of the class served as a guide,and the teacher assigned two small activities to the students.The first activity is finding a students photos in the computer photo album,and the second is finding the
155、 target object in 12 images.The former is a task that a human cannot complete within a specified time;hence,it leads to the second part,explaining and discussing the knowledge of face clustering and face recognition.The third part is experimental exploration.The teacher let students,according to spe
156、cific steps in the SenseStudy AI experiment platform with Python programming experiment,experimental record data,through practical experience how is about three weeks.use face clustering,and after the end of the experiment,to search engines and logistics robots as an example,about real-life applicat
157、ion of is about three weeks.Different from the traditional introduction with problems into the class,they are discarded when their functions are completed,or after completing knowledge learning,teachers ask carefully designed questions to guide student discussion.In problem-based learning,all learni
158、ng activities are developed around the initial problem.The acquisition of individual knowledge is implicit in the process of problem solving.Subject knowledge is the assistant serving problem-solving.Therefore,the problem,as the starting point of learning,is related to whether the problem-based lear
159、ning can be effectively implemented.(iii)Experiential teaching Experiential teaching is a teaching methodology that senses,understands,and verifies the teaching content through personal experience31.It contains several basic characteristics:Experience learning is a learning process.Experience learni
160、ng is a continuous process based on experience.Experience learning is the process of conflict resolution in dialectical confrontation.This teaching methodology is also in line with the constructivist concept of learning,that is,learning is the process of active construction of the subject.In experie
161、ntial teaching,individualcognition,emotion,and behavior of learners can correspond to four stages of development:in cognitive aspects,individuals experience goal orientation,knowledge acceptance,knowledge internalization,and knowledge transference.In emotional aspects,individuals experience emotiona
162、l expectation,emotional emotion,emotional resonance,and emotional sublimation.In behavior,individuals will experience preview perception,activity,introspection communication,and integration32.Case:Zhorai33 34Introduction:Zhorai is a session agent developed by a team at the Mit Media Lab to teach mac
163、hine learning concepts to primary school children.Teaching object:grade 3-5 students.Course content:Students talk to AI dialogue agents on this platform to understand the natural interaction with machines,how machines learn,knowledge and reasoning of how machines represent information,and the social
164、 impact of such agents.Students teach agents about animals,and agents use prior knowledge to guess which ecosystem the animal cbelongs to.In the process,students learn machine perception and reasoning,as well as AI ethics.(iv)Individualized earningEducation,under the influence of AI technology,will
165、surely change.Personalized learning is a broad term,and online learning field expert Mike Sharps defines individualized learning as a combination of learning that combines learners needs,interests,and abilities35.Individualized learning has been the educational goal of the United States for the past
166、 40 years,and strives to provide learning resources and activities designed to suit the needs and abilities of learners.Nowadays,mobile technology and is about three weeks.technology can provide a physical and social environment for individualized learning,and the application of technology in educat
167、ion is technology-enhanced learning.With the rapid development of technology andthe continuous enrichment of information resources,learners can continue to learn and gain experience outside the classroom.Machine learning technology can be used to analyze the data generated by learners,provide teachi
168、ng processes that match their learning Figure 4:Screenshot of the teaching process of the facial recognition course International Report of Supportive Environment on AI Courses in K-12-K-12 AI Textbook and Teaching Methodology15K-12 AI Textbook and Teaching Methodology-International Report of Suppor
169、tive Environment on AI Courses in K-12performance,and push different learning content according to different situations.In individualized learning,adaptive systems are widely used.They can adapt to students behavior and inferred knowledge through machine learning and deep learning technologies,inter
170、vene in learning at an appropriate time,and give effective tips and support.And the adaptive system was designed not to replace teachers but to improve teaching.Since the 1860s,most schools in the United States began using personal learning.Amazon in the United States has deep learning-based image r
171、ecognition,mobile technology,and AI technology that can support students individualized learning in face recognition and emotion analysis;intelligent voice and chatbots can answer difficult questions.Tencent in China also has several achievements in the field of AI.For instance,Tencent has now set u
172、p an AI lab;it also has a clouds big data processing capabilities,which can perform intelligent grading,homework intelligent correction technology,voice recognition technology,and face recognition check-in.In China,schools should be better able to support personalized learning if they can work with
173、companies and establish good partnerships.(v)Dual-Teacher classThe most primitive form of dual-teacher classroom is to organize excellent teachers to record teaching resource discs,distribute them to rural primary and secondary schools,and partially replace the teaching tasks of on-site teachers.Wit
174、h the improvements in science and technology,the current dual-teacher classroom refers to a new classroom model in which the AI education robot and teachers jointly undertake the teaching work in the classroom.Under this teaching style,the AI education robot undertakes part of teachers teaching task
175、s and provides personalized learning services.The lesson objective of a middle school math class is to help students understand the concept of unary quadratic and can skillfully turn the equation into general forms.When analyzing the learners,the teachers provide an overall understanding of the stud
176、ents in the class according to the learning data provided by the AI education robot and classify the data systematically.By using the knowledge graph technology,the knowledge points related to the unary quadratic equation are presented.In terms of presenting the learning content,teachers choose to i
177、mport the equation into a scene and control the AI education robot through voice to project the video.The mobile terminal lets students transmit their questions to the AI education robot.Then,the robot will choose different ways to answer the students,and for the questions that cannot be answered,th
178、e robot will send the questions back to the teachers.At the end of the class,the AI education robot will provide an evaluation report for each student in the class,helping teachers understand each learners learning progress36.(vi)Intelligent classsThe value of using AI in primary and secondary educa
179、tion technology is reflected more centralized in the deep integration with big data technology.Relevant organizations in various regions focus on building basic education databases and trying to build the analysis and application model of education data.The application scenarios of education data ar
180、e mainly reflected in two dimensions of teaching and learning.Based on data,students learning status can be recorded entirely,learning platforms can pus customized content,and teachers can then teach in a more targeted manner.A third-grade classroom at the Hai Bin Primary School in Chengdu,China,uti
181、lized the iFlytek English Listening and Speaking Quality Testing System.The English listening ability assessment is divided into two types of questions:single sentence reading aloud and statement reading aloud,and students wear headphones to answer the questions.Teachers can monitor each computer in
182、 real time in the background to understand the students answering status and ensure fair testing.At the end of the assessment,the system automatically generates class reports and individual student reports,and teachers as well as students can view their individual results and the overall class situa
183、tion on the platform,introducing automatic computer scoring into the assessment.At Hai Bin Primary School,the intelligent voice evaluation system helps English listening and speaking exams and plays an auxiliary role in daily teaching.For instance,if the teacher organizes English speaking activities
184、 in the classroom,the system can help evaluate students oral English in real time,diagnoses speech problems,and has a one-click dictation function set up to reduce the teachers burden.The generated evaluation report can pinpoint students weak points and utilize better materials to help students lear
185、n.Figure 5:Double-teacher class in a middle schoolFigure 6:Students use the intelligent language evaluation system164.K-12 AI Algorithms and Algorithm and DatasetThis chapter provides a brief description of the algorithms for core AI technologies,and screens the open datasets currently suitable for
186、K-12,recommending some datasets.4.1 An AI algorithm suitable for K-12Algorithm,originally algorism,means Arabic numbers,while the English name Algorithm,from the 9th-century Persian mathematician al-Khwarizmi,evolved into algorithm in the 18th century.The Euclidean algorithm is considered as the fir
187、st algorithm in history.The base layer of AI can be divided again according to the algorithm,computing power and data37.The algorithm level includes supervised learning,unsupervised learning,reinforcement learning,transfer learning,and deep learning38.The computing power level includes AI chip and A
188、I computing architecture;the data level includes data processing,data storage,data mining,etc.The application layer can be divided into computer vision,voice processing and natural language processing according to the algorithm use.Computer vision includes image recognition,visual recognition,and vi
189、deo recognition.Voice processing includes speech synthesis,speech recognition and voiceprint Recognition.Natural language processing includes semantic recognition,text proofreading,machine translation,natural language generation,etc.International Report of Supportive Environment on AI Courses in K-1
190、2-K-12 AI Algorithm and DatasetDataArtificialIntelligenceApplicationlayerBase LayerAlgorithmComputing PowerImage RecognitionSpeech SynthesisSemantic RecognitionSupervised LearningData MiningText ProofreadingUnsupervised LearningData ProcessingVideo RecognitionSpeech RecognitionMachine TranslationRei
191、nforcement LearningData StorageLocal GPUVoiceprint RecognitionNatural Language GenerationDeep Learning Cloud GPUComputer Vision ProcessingVoice ProcessingNatural Language ProcessingFigure 7:Core technologies of artificial intelligence17K-12 AI Algorithm and Dataset-International Report of Supportive
192、 Environment on AI Courses in K-12 Example of the K-12 AI algorithm Each AI technology is complex and,for example,machine learning involves both mathematical foundations like statistics,information theory,and cybernetic theory,and other non-mathematical knowledge.These technologies are very professi
193、onal.For K-12 students,they do not need to master a certain complex algorithm in detail,but only need to understand the basic principles,application areas,use methods,etc.Therefore,we selected two typical AI algorithms as the examples.FACIAL RECOGNITION TECHNOLOGY(i)What is facical recognition?Facia
194、l recognition is a technique used to verify or identify individuals by processing a video frame or a digital image of a visible individuals face.Since computerized facial recognition involves measuring the physiological features of humans,facial recognition systems are classified as biometric39.Faci
195、al recognition techniques have several different approaches,but they often compare facial features in images to those included in the database40.Facial recognition systems have been more widely used in smartphones and other forms of technologies(such as robots).For example,police officers can use th
196、is technology to identify a criminal suspect;or some of its applications include automatic image index,video surveillance,human-computer interaction,etc.(ii)What algorithm does facial recognition use?Facial recognition algorithms are an essential component of any facial detection and recognition sys
197、tem or software.Experts divide these algorithms into two central approaches41:the first geometric approach focuses on distinguishing features.The regional feature analysis algorithm is widely used in facial recognition technology,which integrates computer image processing technology and biostatistic
198、s principle,uses computer image processing technology to extract portrait feature points from video,and uses the principle of biostatistics to analyze and establish mathematical models,namely face feature template.Another photographic statistics was used to extract values from images and then compar
199、ed to the template to eliminate variance.In the general direction,the algorithm can also be divided into two more general categories,namely,partial and overall models.The former focuses on facial features and analyzes its spatial parameters and correlations with other features,while the holistic app
200、roach treats the face as the entire unit.The algorithms involved in facial recognition are convolutional neural network(CNN),EIGENFACES,FISHERFACES,PCA,SVM,etc.(iii)Use facial recognition technology in schoolsFacial recognition technology is now being introduced in all aspects of public life.This in
201、cludes the rapid integration of facial recognition and facial testing into K-12 education to address issues such as campus security,automatic registration,and student emotion testing42.In this controversial context,we need to consider applying this technology to a specific educational context.A prom
202、inent educational application of facial recognition technology is campus security.In addition to identifying unauthorized intruders,systems were developed to detect video objects carrying dangerous items into school43.Another application of facial recognition in schools is attendance monitoring,whic
203、h avoids the time waste or omission caused by the teachers roll call in class.In addition,research and development reports in the field say that detecting short facial movements(i.e.,emotional test)can prove whether students are interested in the course,in order to achieve personalized teaching,dete
204、ct their emotions in real-time,respond in time,and encourage students to listen to classes or interact actively.These systems can all help teachers with their teaching activities.DECISION TREE ALGORITHM(i)What is a decision tree algorithm?Decision tree is one of the most widely used and most practic
205、al machine learning algorithms,which involves little mathematics and is very easy to use and interpret44.Decision tree algorithms belong to a family of supervised learning algorithms.Unlike other supervised learning algorithms,the decision tree algorithms can also be used to solve regression and cla
206、ssification problems.The goal of using a decision tree is to create a training model that can predict the level or value of the target variable,by learning simple decision rules inferred from previous data(training data).(ii)Application of the decision tree algorithm in schoolsThe planning and desig
207、n of the student course selection system is particularly complex and involves a variety of algorithms45.At the same time,whether the course selection system is reasonable and science directly affects the teaching quality of the whole course selection and students enthusiasm for elective courses.Deci
208、sion tree algorithm and other techniques are used to uniformly classify student information and comprehensively analyze students hobbies and needs to achieve target data mining.In addition to playing a role in the student course selection system,it can also be used in the administration of universit
209、y educational affairs46.At present,the number of colleges and higher learning is increasing.In order to promote the reasonable and scientific development of schools,the classification rules can provide some help for academic administrators,so the decision tree algorithm plays a certain role in these
210、 two aspects.18 Algorithmic tools and frameworkAt present,the basic algorithm of AI has been relatively mature,and major enterprises have made great efforts to build an algorithm model tool library,and encapsulate it as a software framework for developers to use.It can be said that the software fram
211、ework is the engineering implementation of the algorithm.The software framework of an enterprise is implemented in two forms:closed source and open source.The following listed programming tools and frameworks are all open source projects and can be free to use.NAMEINTRODUCTION URLScratchScratch is a
212、 free program developed by the MIT Media Lab Lifetime Kindergarten Group,designed mainly for children aged 8 to 16,but people of all age groups can use Scratch to create and share.Younger children can try ScratchJr,a simplified version of Scratch for children ages 5 to 7.https:/appinventor.mit.edu/A
213、pp InventorApp Inventor is a fully online developed Android programming environment,discarding the complex program code and using the building block stacking method to complete the Android programs.It supports Lego NXT robot,suitable for Android beginners or robot primary development.https:/www.pyth
214、on.org/PythonPython is one of the most widely used programming languages in the AI space,and can be used seamlessly with data structures and other commonly used AI algorithms.Python has a large library of tools,making others have great advantages in AI programming.https:/www.python.org/NAMEINTRODUCT
215、ION URLTensor FlowTensorFlow is an open source framework based on data flow programming developed by Google and is widely used for the implementation of various types of AI algorithms.TensorFlow supports multiple programming languages like C,Python,and Java,allowing computing on any CPU or GPU,wheth
216、er it be a desktop,server,or mobile device.https:/ Cognitive Toolkit(CNTK)CNTK is a free,open-source business-level AI development tool framework developed by Microsoft,Inc.It has a variety of AI classical algorithms that can quickly build and train some popular deep learning systems.CNTK was develo
217、ped using C+,but is often called using the Python program.https:/ is an open-source Python machine learning library developed by the Facebook Artificial Intelligence Research Institute,providing powerful tensor computation based on GPU acceleration,while supporting dynamic neural networks.https:/pyt
218、orch.org/Flying Paddle(PaddlePaddle)Based on Baidus deep learning technology research and business application.It integrates deep learning core training and reasoning framework,basic model library,end-to-end development suite and rich tool components.Airpaddle helps developers quickly realize AI ide
219、as and quickly launch AI business.Helps more and more industries to complete AI empowerment and realize industrial intelligent upgrading.https:/ is an open source artificial neural network library written in Python that can be used as a high-level API for Tensorflow,Microsoft-CNTK,and Theano for dee
220、p learning model design,debugging,evaluation,application,and visualization.https:/keras.io/TABLE 6:THE AI PROGRAMMING TOOLTABLE 7:THE ARTIFICIAL INTELLIGENCE DEVELOPMENT FRAMEWORK International Report of Supportive Environment on AI Courses in K-12-K-12 AI Algorithm and Dataset 19K-12 AI Algorithm a
221、nd Dataset-International Report of Supportive Environment on AI Courses in K-124.2 Challenges faced in the teaching of K-12 AI algorithm4.3 The Outlook on Algorithm Teaching in K-12 AI Education4.4 Dataset suitable for K-12 AI educationBy summarizing the existing standards,books and platform of algo
222、rithm teaching in K-12 AI education,the challenges facing algorithm teaching are:(i)Teaching standards have yet to be harmonized across countries.(ii)Difficulty in quantifying and assessing the sequence and algorithmic difficulty of the teaching content.(iii)There is room for improvement in the Inte
223、rnet platform learning environment.(iv)There are no guidelines or examples of the fit between the teaching of algorithms and the teaching of basic subjects.With the release of a series of AI standards(e.g.,AI4K1248,Chinese primary and secondary school AI curriculum development standards49,etc.),it c
224、an be seen that countries are to launch an AI education aimed at developing national guidelines for teaching AI content at the K-12 level,providing a resource catalog for teachers,and establishing a K-12 AI resource development community.As for the teaching of algorithms for them,the learning of bas
225、ic principles should be the focus of teaching,and the learning of algorithms needs the support of mathematics,information technology and other disciplines.Therefore,the future for the teaching of algorithms in K-12 AI should be extended from the basic disciplines,and the learning of algorithms shoul
226、d correspond to the learning of basic disciplines,so that they can support each other moreeasily and better.For the teaching of algorithms in AI,an AI experience course can be conducted to familiarize primary and secondary school students with AI technologies and product applications from a basic pe
227、rception level.At the same time,at the primary and secondary education level,teachers are allowed to infuse design thinking into their teaching as a way to cultivate students innovative thinking and creative awareness,thus AI data plays an irreplaceable role in the commercialization of AI.When an al
228、gorithm model is designed,a large amount of annotated data is needed to train the machine,thus making the machine more intelligent so it can be used in actual application scenarios.If the algorithm is expected to further improve the performance,more refined data needs to be trained and constantly it
229、erated.At the same time,there are still unsolved risks in the development of intelligent education:intelligent education governance risk,intelligent education technology risk,intelligent education ethics risk,and teacher career substitution risk47.There are also many challenges in the teaching of AI
230、,such as lack of understanding of curriculum value and positioning,limited school hours,lack of relevant teachers and training strength,and inadequate software and hardware environment needed to carry out AI teaching.In addition,most teachers are at a loss for the students with great differences in
231、knowledge reserves.helping them to meet the challenges of the future AI era.A dedicated AI application course will allow students to apply AI in project-based and inquiry-based learning.The AI application course can inspire students and prepare them to enter the AI workplace in the future.By focusin
232、g on the overall project development process,an AI R&D course is set up to analyze real-world problems,to promote algorithmic learning50.AI has now become the integration of the current information revolution,and the algorithm realization of AI needs people,and its purpose is also to seek welfare fo
233、r human beings.School education committed to talent training will only become more and more important in the future,which requires us to focus on the future and think deeply about what the mission of school education is and what the focus is.Predictably,AI education should not be the application of
234、AI technology only,is not a simple superposition of technology and education,but should scope beyond AI,it should return to the basics that is to say,education should return to the essential relationship between people and professional education and social functions,not blindly worship information t
235、echnology,and the goal of peoples happiness as the pursuit of education51.20 The datasets of the representative AI technical fieldThe following datasets are some of the typical datasets commonly used for different AI technology fields in youth AI teaching.Representative professional AI datasetFor di
236、fferent application fields,this report collects some relatively representative and typical datasets for your reference in conducting research or teaching in the corresponding fields.Representative professional AI datasetFor different application fields,this white paper collects some relatively repre
237、sentative and typical datasets for your reference in conducting research or teaching in the corresponding fields.Test and adopt emerging AI technologies and tools for ensuring teachers and learners data privacy protection and data security.Support robust and long-term study of deeper issues of ethic
238、s in AI,ensuring AI is used for good and preventing its harmful applications.Develop comprehensive data protection laws and regulatory frameworks to guarantee the ethical,non-discriminatory,equitable,transparent and auditable use and reuse of learners data.from Article 29 of the Beijing Consensus In
239、ternational Report of Supportive Environment on AI Courses in K-12-K-12 AI Algorithm and Dataset APPLICATION FIELDNAMEINTRODUCTION URLHandwritten recognitionMNISTMNIST is a picture dataset of handwritten figures,with a total of 250 handwritten figures from different people.MNIST is divided into two
240、parts:training set and test set,in which the training set contains 60,000 images and labels,while the test set contains 10,000 images and labels.http:/ recognitionImageNetThe Imagenet dataset is used in the deep learning image field,which can be used for image classification,positioning,detection an
241、d other research work.The dataset has more than 14 million images covering more than 20,000 categories,with more than a million images with explicit category annotation and the location of objects in the image.https:/image-net.org/Image recognitionCIFAR-10The CIFAR-10 contains 10 categories,60,000 t
242、raining images,and the color image size:32x32,10,000 test images.The CIFAR-100 contains 100 classes with 600 images,500 for training and 100 for testing;the 100 classes form 20 superclasses.Image categories are all clearly marked.CIFAR is a very good small and medium-scale dataset for image classifi
243、cation algorithm testing.http:/www.cs.toronto.edu/kriz/cifar.htmlSpeech recognitionCHIMECHIME is a dataset for Speech Recognition Challenge(CHiME Speech Separation and Recognition Challenge).The dataset contains three parts:training set,developer machine,and test set,and each copy includes data from
244、 multiple speakers in different noise environments.http:/spandh.dcs.shef.ac.uk/chime_challenge/index.htmlNatural language processingSQuADThe Stanford Q&A Answer Dataset(SQuAD)is a new reading comprehension dataset consisting of questions extracted from Wikipedia,and each answer is a piece of text in
245、 the corresponding paragraph.Over 100,000 Q&A pairs in over 500 articles.https:/rajpurkar.github.io/SQuAD-explorer/TABLE 8:DATASETS OF THE REPRESENTATIVE AI TECHNICAL FIELDS21K-12 AI Algorithm and Dataset-International Report of Supportive Environment on AI Courses in K-12APPLICATION FIELDNAMEINTROD
246、UCTION URLAgriculturelemonThe lemon data contains 2,690 annotated lemon images(1,056 x 1056 pixels)containing three subsets of training,validation,and testing.Each lemon image was taken at different angles and taken multiple times under different lighting conditions.An individual fruit has generated
247、 approximately 100 images.https:/ Cancer Cell Line Encyclopedia Dataset CCLECCLE covers more than 1,000 human cancer cell lines covering over thirty tissue sources for large-scale deep sequencing,integrating genetic information such as DNA mutations,gene expression,and chromosomal copy number.This d
248、atabase has become one of the standard reference databases for cancer genomics.https:/sites.broadinstitute.org/ccleClimateThe Dutch weather data platform,KNMIThe Dutch Weather Data Platform(KNMI)has about 247 datasets containing 71 climate datasets,78 precipitation datasets,7 seismic and acoustic da
249、tasets,13 sunlight and radiation datasets,33 temperature datasets,7 weather forecast datasets,and 38 wind datasets.https:/dataplatform.knmi.nl/Complex networkStanford GraphBaseThe Stanford GraphBase is a collection of programs and datasets that generate and manipulate graphics and networks.This pack
250、age is the work of Donald Knuth at Stanford University,and the latest version of this software is always provided by an anonymous ftp from the Stanford Department of Computer Science.https:/www3.cs.stonybrook.edu/algorith/implement/graphbase/implement.shtmlMachine learningUCI machine learning librar
251、yThe UCI machine learning library contains 588 machine learning-related datasets of various types,of which there are classical ones like Iris,and it also contains new added datasets such as air quality and GPS trajectories.https:/archive.ics.uci.edu/ml/index.phpComputer networkCAIDA data setCAIDA co
252、llects network data from dimensions such as geography and network topology and provides this data to the research community as much as possible.This dataset can analyze two categories of network data collected and being collected.https:/www.caida.org/catalog/datasets/overview/Network securityCCCS-CI
253、C-AndMal-2020CCCS-CIC-AndMal-2020 is a new comprehensive and massive robotic malware dataset consisting of 20 0,000 benign and 20 0,000 malware samples that total 40 0,000 Android applications,containing 14 prominent malware categories and 191 prominent malware.https:/www.unb.ca/cic/datasets/andmal2
254、020.htmlGeo-information systemWorld continental,national,and city datasetsThe data contains all 7 continents,250 countries/regions,4k subdivisions(provinces,states,etc.),and over 127,000 cities.It shows all the languages used in each country/region,and the geographical shape of each country.All data
255、 are provided in the JSON format and can be retrieved and managed through the API.https:/ collects,analyzes,and disseminates data and information on water,water and agricultural water management by country,with a focus on irrigation agriculture in Africa,Asia,Latin America and the Caribbean.http:/ww
256、w.fao.org/aquastat/statistics/query/index.html;jsessionid=84948F118DA73E55A555F8AF0642005FTABLE 9:DATASETS OF THE REPRESENTATIVE SPECIALTIES22 International Report of Supportive Environment on AI Courses in K-12-K-12 AI Algorithm and Dataset APPLICATION FIELDNAMEINTRODUCTION URLEconomicsEconomic dat
257、a from all countries around the worldEconomic data from the American Economic Society collects American macroeconomic data,as well as economic data from other countries.It can provide a growing variety of economic data for students,teachers,and professionals.https:/www.aeaweb.org/resources/dataEduca
258、tionThe New York State Education datasetThe New York State Education Dataset collected New York State education data from 1999 to 2016,containing a transcript database,enrollment database,graduation rate database,etc.https:/data.nysed.gov/downloads.phpFinanceThe FAANG stock datasetThe FAANG stock da
259、taset contains stock data when FAANG Corporation began trading.These data includes data on the opening price,closing price,sales of these stocks.https:/ Open dataset CODCOD is an open access database that collects crystal structures of organic,inorganic,metal organic compounds and minerals,but not b
260、iopolymers.The COD database currently has more than 480,000 pieces of data.http:/ cultureThe CRICSHEET Cricket DatasetThe CRICSHEET Cricket Data provides structured score data for more than 10,000 cricket games.https:/cricsheet.org/TrafficThe Freight Analysis Framework,FAFThe Freight Analysis Framew
261、ork(FAF)integrates data from a variety of sources to provide a comprehensive understanding of freight flow through various means of transportation between US states and major metropolitan areas.The FAF included data from agriculture,mining,utilities,construction,services,and other sectors.https:/ops
262、.fhwa.dot.gov/freight/freight_analysis/faf/index.htmCONTINUED FROMTABLE 9:DATASETS OF THE REPRESENTATIVE SPECIALTIES23AI Laboratory for K-12-International Report of Supportive Environment on AI Courses in K-125.AI Laboratory for K-12This chapter briefly describes the AI experimental environment,whic
263、h is divided into three types:economy type,standard type,and advanced type.Different types of experimental environments correspond to different learning groups,and you can choose the most suitable learning(teaching)platform.Experimental environment is necessary for learning AI,through which learners
264、 can practice the theoretical knowledge learned.Many students cannot innovate in the laboratory if they have not practiced in the laboratory.AI teaching laboratory is an important infrastructure for developing AI education.The AI labs focus on learning and using new technologies to solve real-life p
265、roblems.AI labs can inspire students to prepare for the fourth industrial revolution in AI,robotics,and automation technologies.On the basis of fully considering the school economic situation and teaching positioning differences,the research group proposed the classification of AI education laborato
266、ry as shown in Figure 8.It is suitable for some undeveloped and backward countries or regions.This method is a simple and low-cost solution based on open and free economic resources.It is suitable for some impoverished and backward countries or regions.It is a simple and low-cost solution based on o
267、pen and free economic resources.The upgrade on the basis of the standard type has two application directions:research and teaching.The research direction has fewer projects and fewer staff,but the most sophisticated equipment are used.The teaching direction requires more faculty to support teaching
268、activities because of the large number of staff and projects,so there is a big difference between the two.Economy typeStandard typeAdvanced typeFigure 8:The AI Laboratory Classification24 International Report of Supportive Environment on AI Courses in K-12-AI Laboratory for K-125.1 Classification an
269、d characteristics of AI laboratoriesDifferent types of experimental environments correspond to different learning groups of learners,one can choose the learning(teaching)platform that best suits their needs.An experimental environment is a necessary condition for learning AI,through which learners c
270、an put the theoretical knowledge they have learned into practice52.Because many students have not practiced in a lab,it is difficult to gain new technological innovations from theoretical courses.With an environment for learning AI,the number of AI talents will continue to grow in the future.(i)Econ
271、omy type of AI labIn 2018,Africa,a most dedicated teacher on the Internet fire,because of poverty,the computer in their side is almost an inaccessible thing,then their students how to take computer classes?The teacher had a brainstorm,he took chalk hand-drawn computer page on the blackboard,so as to
272、 teach the students.Many schools do not have robust lab facilities and rely mainly on boring textbooks to teach AI,which can make learning much less effective.Most of the economical labs currently existed are free and open,which include providing a free programming platform,GPU computing power and s
273、ome courses that all people can use or teach for free.The characteristics of this kind of lab are:anytime,anywhere,flexible,open,and able to accommodate all people who want to learn AI technology.Some platforms are free to use,which is more suitable for those who are living in poverty.Economical lab
274、s log in to AI cloud services via the schools computer room WEB and conduct experimental teaching using the free AI computing power and services provided by these cloud services.Currently,such cloud services are provided by Baidu,Ali,Google and other companies.(ii)Standard type of AI labAt present,t
275、here are six major research directions in AI technology,among which computer vision,natural language processing,robotics and machine learning are relatively popular,and more and more products are being applied in related fields.Standard laboratory characteristics:enough to support the personal works
276、tation equipment,can be equipped with their own equipment according to different project needs,set up suitable for their own workstation.This laboratory does not cost too much and is suitable for an individual laboratory.(iii)Advanced type of AI labAI is the most important development field of the f
277、uture,and the demand for students or users is growing.And many students who have not practiced in the lab and have only taken theoretical courses alone will have a hard time using those theories to innovate technologically in the workplace.Creating an AI lab is a huge challenge for schools because i
278、t is not clear what equipment is needed for each area of AI53.And thus,we need to analyze the different areas of AI and then make a professional recommendation for schools.The features of the advanced type of AI lab include the unified of the school,the equipment required for professional and compre
279、hensive,allowing students to thoroughly learn professional knowledge and cutting-edge AI technology.However,the cost for this kind of lab is the most expensive among the three other labs,so it is more suitable for large institutions,such as schools,to build.Figure 9:Hand-painted computer by African
280、teachers25AI Laboratory for K-12-International Report of Supportive Environment on AI Courses in K-12Figure 10:Xiongan AI Education Laboratory of Xiongan New AreaFigure 11:The I N-MaC Design and Innovation Studio IN-MaC Design and Innovation Studio54On March 13,2020,the Indiana Next Generation Manuf
281、acturingCompetitiveness Center(IN-MaC)and Honda Manufacturing Cloud for Youth Digital Learning Space for Rural StudentsCloud for Youth is a new initiative aimed at building a new learning space for rural students.The initiative is founded with participation from governments,companies,and civil socie
282、ties to lead the betterment of digital education for rural students.Leveraging cloud technologies,the initiative provides a virtual personal desktop,which integrates classroom management capabilities and provides digital capabilities including resource storage and digital laboratory.This is expected
283、 to create a conducive learning environment for the IT education of primary and secondary students,providing students with access to AI-powered experiments and after-school learning programs.As the provider of cloud-based services and hardware for the initiative,Alibaba Cloud will provide 5 years wo
284、rth of free deployment and technical services.Indiana LLC(HMIN)debuted the latest IN-MaC Design and Innovation Studio.Located at the HMIN facility in Greensboro,this lab offers hands-on experience with modules such as additive printers,robotics design and AI coding,as well as engineering and science
285、 learning.In addition,it includes virtual reality experience stations and equipment for STEM education.With the goal of enhancing middle and high school students exploration of design thinking,problem solving,technology and innovation skills,the lab enables students from southeastern Indiana,USA to
286、discover and explore the next generation of manufacturing.The strength of this lab is that,on the one hand,it allows students to put into practice the theoretical knowledge learned in STEM courses and develop their problem-solving skills and interest;on the other hand,it brings a real work environme
287、nt into the school learning environment,providing a unique way of bridging the gap between industry and education.In this way,students are provided with relevant experiences that help develop their IT skills,employability skills,form career concepts and develop new ways of thinking.5.2 K-12 AI lab e
288、xample The Artificial Intelligence Education Laboratory in the Xiongan New AreaThe laboratory is established by Baidu Education,Smart Learning Insiture of Beijing Normal University and Baiyangdian senior high school.This lab could support the learning parctice of AI,robot,big data,internet of thing
289、and brain interaction,which is smart learning solution integrating curriculum resource,teaching platform,practical teaching and virtual experience.This lab will serve the AI teaching,STEM education and teacher training at shcool.The AI curriculum is divided into three main categories according to th
290、e K-12 age group.The elementary school level focuses on computational thinking development,aided by the perception and experience of artificial ntelligence.The middle school level focuses on advanced computational thinking development,aided by an initial understanding of the AI implementation knowle
291、dge system.The high school level focuses on thematic AI technical understanding,aided by the frontiers of AI educational development and application.The learning focus and specific measures for these three stages are introduced in detail,and the three stages comprehensively cover basic education.Sol
292、utions are also proposed for the training of K-12 teachers through online and in-person training.In this way,teachers can understand the history,current situation,methods,and significance of AI technology and understand the teaching.methods and outcome evaluation criteria of AI courses.Teacher train
293、ing ensures that students can master the cognition of AI technology,understand the definition of AI sub-disciplines,and grasp the principles of AI technology through practical activities.Educators are required to pass training assessments before starting teaching.26Figure 12:“Cloud for Youth”Digital
294、 Learning SpaceThe learning space is equipped with 4050 all-in-one lightweight devices that provide access to the virtual CPUs,GPUs,memory,and disks deployed on Alibaba Cloud.This initiative solves long-standing issues where rural schools lack access to modern computing technology and staff to manag
295、e and maintain IT equipment.Through this initiative,rural students are provided with equal access to digital learning and experiment spaces as students in urban schools.Cloud for Youth promotes equitable access to technology for students in rural areas.Teaching resources and tools will be deployed o
296、n the cloud and every teacher and student will be able to log on to their own personal desktops and access their resources at anytime,anywhere,and on any device.Teachers and students no longer have to be bound by the confines of the school and can continue their education from the comfort of their o
297、wn homes.The learning space also creates a contextualized teaching environment for AI-powered classes.Through cloud networking,dynamic distribution,and peer-to-peer management,students can collaborate with each other in learning and can interact with the teacher during classes.This way,students can
298、gain a better understanding of the importance of digital technologies and cultivate creative thinking through collaboration and interaction.International Report of Supportive Environment on AI Courses in K-12-AI Laboratory for K-1227Conclusion and Recommendation-International Report of Supportive En
299、vironment on AI Courses in K-126.Conclusion and RecommendationAI is increasingly adding its important application value in many fields with the advent of the intelligent era.However,AI education still faces many practical problems in basic education,such as fuzzy curriculum orientation,differentiate
300、d teaching content,and complex curriculum system and resources.Looking forward,we will find that the development of K-12 AI education will still go through a long stage of development.6.1 The Conclusion on K-12 AI Education The development on AI open education resources will lead to a fair education
301、al resource distributionThe AI OER is the teaching,learning,and research material for any media(digital or other media)located in the public domain or issued with open licenses.The license allows free access,use,adaptation,and redistribution without user restrictions.Although the current OER granted
302、 by some institutions or famous universities may help to create fair educational resource distribution,which still lacks suitability among the users.Teaching materials and methodologies are indispensable for developing AI in K-12Textbook and teaching methodologies are regarded as complementary to ea
303、ch other.Textbooks are the main carrier of teaching content,in they serve as a guide to help learners meet their education goals.AI textbooks for kindergarten and primary school centralize the popularization of AI knowledge and the cultivation of student skills.In the meantime,AI textbooks for junio
304、r high school focus on the cultivation of AI techniques and engineering thinking.For high school students,AI textbooks emphasize fostering students ability and mindset through hands-on practice to understand and apply AI.This teaching type through practice in AI education has appeared to be successf
305、ul.Algorithmic literacy is an essential ability for K-12 educators Three elements of AI are calculating skills,algorithms,and data.Although these skills are required and essential in AI education,many K-12 educators lack algorithmic literacy.Even if they possess the skill of algorithmic literacy,it
306、is often included in the curriculum of computer science instead of AI education.To help develop students algorithmic literacy55,we should first put attention to the educators own algorithmic literacy55.It is recommended to first ensuring that AI educators possess sufficient algorithmic literacy skil
307、ls,then utilize project-based or case-based learning to teach K-12 students.The AI laboratory for education is a key factor in the development of K-12 AI teachingSuppose we separate the AI teaching laboratory into three scenarios.In that case,the economy type is to provide free resources in less dev
308、eloped countries like Paddle OCR by Baidu,which gives free online coding environment,GPU algorithm,and algorithm and open data.The second scenario is the standard type,suitable for setting up personal AI workstations;it allows the learners to open up deeper learning,data science,reinforcement learni
309、ng,and AI all-in-one personal workstations.The third scenario is the advancement type.This third type of scenario is suitable when school builds their laboratory for learning.This lab can accommodate more students to do activities,present models,corroborate,and be creative in this environment.This l
310、aboratory will provide students an opportunity to participate in practical learning activities in AI professionalsworking environment.The educational policy for AI needs to be future improved to achieve SDG.Referring to the Beijing Consensus,when developing AI education,we should recognize the multi
311、disciplinary nature of AI,ensuring the coordination between educational AI and public policy,plan and govern educational AI policies through all government participation,cross-departmental integration,and multiple collaboration.Identify strategic priority areas of policy based on local challenges in
312、 achieving SDG and its specific goals and other SDG efforts.From the perspective of lifelong learning,to complete and improve AI education,we should plan and formulate the whole system with AI educational strategy by integrating education policies.28 The AI policy and safety are critical for owners
313、to avoid legal and ethical issues.When governments and education departments support AI development,AI applications may have different types of biases.These biases are inherited in the input and output of AI data or the algorithms used in the system.We should recognize the dilemma between open acces
314、s and privacy protection on data,the legal issues and ethical risks associated with data ownership,data privacy,and the availability of data serving increases the public interest.In addition,we should adopt principles of ethics and privacy and utilize design to ensure security when building AI.6.2 K
315、-12 AI Education RecommendationWith significant progress over the past decade and during the COVID-19 pandemic,AI adoption has accelerated worldwide and across different sectors.The global AI market is expected to reach a staggering$312.4 billion by 2027.Therefore,governments worldwide have prioriti
316、zed AI,launching national strategies,encouraging research and development,promoting regulatory reform,and increasing the talent pool.The talent shortage has been highlighted as a vital issue in AI policy.Major decision-makers in government,academia,and private enterprise sectors operate by initiatin
317、g and increasing support for AI education and esearch,particularly during the educational phase of K-12,as they view AI as primarily an investment in economic competitiveness.However,there is a general lack of information about implementing AI into K-12 education curricula.Some countries include AI
318、education in national-level plans to introduce AI education systematically,while others still rely on nitiatives from local or university organizations.This divergence suggests that creating opportunities for international discussions on K-12 AI education is crucial,sharing each others good practice
319、s and identifying solutions to common challenges.According to the research and analysis of this support report,the author team provides the following suggestions for government,private sector and civil society:governments must invest in K-12 AI education,provide equal access to basic knowledge and s
320、kills,and encourage private sector and private society participation,but the government should ensure the quality and consistency of AI education.Finally,educators should be encouraged to seek breakthroughs in education.AI should not be the only tool for professional computer teachers.Any subject ca
321、n seek the assistance of AIED.International Report of Supportive Environment on AI Courses in K-12-Conclusion and Recommendation29References-International Report of Supportive Environment on AI Courses in K-127.ReferencesMichele.K-12 Artificial Intelligence Market Set to Explode in U.S.and Worldwide
322、 by 2024.2018.https:/marketbrief.edweek.org/marketplace-k-12/k-12-artificial-intelligence-market-set-explode-u-s-worldwide-2024/Touretzky D S,Gardner-McCune C,Martin F,et al.K-12 guidelines for artificial intelligence:what students should knowC/Proc.of the ISTE Conference.2019.Wu Fenghua,Yu Jiawei.D
323、oes AI create value?revisit of labor based on the three-dimensional analysis framework J.Humanities Magazine,2020(09):36-45.DOI:10.15895/ki.rwzz.2020.09.005.Tedre M,Toivonen T,Kahila J,et al.Teaching Machine Learning in K12 Classroom:Pedagogical and Technological Trajectories for Artificial Intellig
324、ence EducationJ.IEEE Access,2021,9:.United Nations Educational,Scientific and Cultural Organization.UNESCO releases report on the mapping of K-12 Artificial Intelligence curriculaEB/OL.2022-02-05.https:/www.unesco.org/en/articles/unesco-releases-report-mapping-k-12-artificial-intelligen
325、ce-curricula.Chen L,Chen P,Lin Z.Artificial intelligence in education:A reviewJ.Ieee Access,2020,8:75264-75278.Jiang Xin,Zhu Hongyan,Hong Ming.Research on Artificial Intelligence in Education:Tracing and Review J.China Distance Education,2020(2):13.Deloitte.The Global Education Intelligent Developme
326、nt Report.2019.https:/ of Education,PRC Republic of China.Information Technology Curriculum Standards for General High Schools M.Peoples Education Press,2017.UNESCO.The Beijing Consensus on Artificial intelligence and Education.2019.https:/unesdoc.unesco.org/ark:/48223/pf0000368303Zhou Shaojin,Wang
327、Fan.The Logical Thinking of K-12 Artificial Intelligence Education:The Road of Student Intelligence Generation and Discussion on K-12 Artificial Intelligence Textbook J.Modern Education Technology,2019,29(04):12-18.Pedro F,Subosa M,Rivas A,et al.Artificial intelligence in education:Challenges and op
328、portunities for sustainable developmentJ.2019.Zhan Zehui,Zhong Baichang.What high school AI education should teach and how to teach based on the content analysis of four AI Preliminary textbooks J.Electrical Education Research,2020,41(6):8.Yu Ping,Zhu Zhiting.Copyright and Access License Research fo
329、r Open Educational Resources J.Open Education Research,2009,15(6):6.Brooks C F.Toward hybridised faculty development for the twentyfirst century:blending online communities of practice and facetoface meetings in instructional and professional support programmesJ.Innovations in Education and Teaching
330、 International,2010,47(3):261-270.Plasencia A,Shichkina Y,Surez I,et al.Open source robotic simulators platforms for teaching deep reinforcement learning algorithmsJ.Procedia Computer Science,2019,150:162-170.De Ridder D.Artificial intelligence in the lab:ask not what your computer can do for youJ.M
331、icrobial biotechnology,2019,12(1):38-40.Indiana University of PA.OER REPOSITORIES AND PROVIDERS.https:/libraryguides.lib.iup.edu/oer/finding_oerTlili A,Zhang J,Papamitsiou Z,et al.Towards utilising emerging technologies to address the challenges of using Open Educational Resources:a vision of the fu
332、tureJ.Educational Technology Research and Development,2021,69(2):515-532.https:/zh.wikipedia.org/wiki/%E9%96%8B%E6%94%BE%E6%95%99%E8%82%B2%E8%B3%87%E6%BA%90Luo T,Hostetler K,Freeman C,et al.The power of open:benefits,barriers,and strategies for integration of open educational resourcesJ.Open Learnin
333、g:The Journal of Open,Distance and e-Learning,2020,35(2):140-158.Li Shengping,Li Shuang,Sun Hongtao.The Online Practice Community Environmental Design Framework for Promoting Teacher Peer Mutual Assistance J.China Distance Education,2010(11):23-28.DOI:10.13541/ki.chinade.2010.11.003.Yu Y,Chen Y.Design and development of high school artificial intelligence textbook based on computational thinkingJ.