In formal education we usually understand a learning process that is structured and systematic, is guided by educated teachers, takes place in class-rooms and institutions like schools or universities, and there are standards and certifications that are ruled by governmental bodies. In contrast, informal education is usually not bound to institutions or given structures and could happen at various places and in many forms.
Recently we observed big shift from formal to informal education. Due to recent advances in virtualization of teaching, numerous organizations like Udacity, Coursera or Udemy, to name a few, emerged providing informal education. Some of them are affiliated with top universities and provide high-end courses. This is a major transition that traditional educators like schools or universities need to recognize. A positive effect is that learning and technology will be in some way democratized.
When we look at AI education for K-12 we see both pillars. On one hand, there are numerous grass-root initiatives that offer material, courses, and training for AI in different quality, density, and quantity. Usually such activities move faster and are able to react to new topics, trends or needs much quicker. On the other hand, governments and supporting institutions and associations realized the necessity to integrate AI education in their national education systems and to provide more structure for AI at K-12 level in terms of curricula, guidelines, or standards. Obviously these activities are slower but gained more attention recently.
One example of formal education in AI for young people is the AI4K12 initiative . The initiative is supported by the National Science Fund (NSF) and the Association for the Advancement of Artificial Intelligence (AAAI) and cooperates with the Computer Science Teachers Association (CSTA) which aims at improving the computer science education in schools by providing corresponding standards. AI4K12 aims at the development of national guidelines for AI education for K-12. Due to the institutional background and targeting teachers as valuable multipliers the activity has clearly a formal aim. An asset of this activity is the development of the standards around the Five Big Ideas in AI (perception, representation and reasoning, learning, natural interaction, and societal impact) . These ideas cover a broad range of topics in AI very naturally. The developed guidelines nicely break down the ideas to topics that should be taught at different age groups and provide concepts and learning activities . Moreover, it provides an online repository for teaching material. Visions of AI4K12 that are shared among many initiatives are (1) to educate citizens to understand AI and allow informed discussions and (2) to educate the AI-literate workforce of the future.
AI Singapore (AISG) is another formal approach to AI education. It was initiated by Singapore’s National Research Foundation and has the overall goal to improve the competitiveness of Singapore in AI. Besides, research and industry education plays a strong role. There are two programs related to AI K12 education. Both follow the train-the-trainer concept in order to improve the skills of teachers and to reach out into schools. The AI4Kids initiative aims at youngsters at the age between 9 and 12 years and focuses on basics in AI with a strong connection to coding . Teachers can obtain certifications from introduction workshops. The initiative has a narrower scope compared to AI4K12, as the topics are more technology-oriented and no general curriculum is envisioned. The AI4Students initiative aims at students at high schools and colleges. This initiative cooperates with the company DataCamp and focuses mainly on big data and data analysis. The initiative acts as a kind of proxy for the existing courses of DataCamp and their learning analytical tools.
After the announcement of the New Generation Artificial Intelligence Development Plan by the Chinese government , a first textbook for AI education was published. The textbook comprises learning activities for a wide grade-band, from primary to secondary school. Details about the content of the textbook can be found in . Yet, it is not sure if and how this activity is integrated into the official curriculum.
South Korea also follows the formal approach to support the competitiveness in AI by providing official guidelines and material. The learning is organized around key areas like understanding AI, AI and data or application of AI and is diversified for the individual grade-bands. In the guidelines there is a strong focus on data science and machine learning with a strong connection to coding as method of learning. A more detailed description of the South Korean initiative can be found in .
Australia follows a different approach to foster AI education for youngsters. Following the general Digital Technologies section of the Australian Curriculum, which defines in general the expectations on acquired skills of students of different grades, AI topics are mapped on that guidelines rather than developed as a standalone AI curriculum. Using the quite open description of national curricula with parts for understanding and building digital systems, concepts from AI like representations, data, or algorithms can easily be integrated. The MOOCS program of the Computer Science Education Research Group (CSER) provides a collection of introduction MOOCS to AI for teachers in primary and secondary schools.Footnote 1 The courses are low-profile and aim to empower teachers to start teaching AI in school. As an introductory course the program has a strong focus on easily accessible topics from the machine learning area.
In the European Union (EU) education is a competence of the individual member states. In the area of AI there is a strong focus of the European Commission towards research and industry. In the White Paper On Artificial Intelligence—A European approach to excellence and trust education is mentioned in the sense of using AI to improve education as well as educate people to understand AI but no detailed plan for AI education K-12 is given. Although there is no development of curricula on the EU level there are initiatives like AI Basis for Schools that is integrated in the European CodeWeek that aim on training teachers in foundations of AI to allow them to integrate AI into their national curricula (to the extend the curricula allow such an integration).
A final example of an initiative to support formal education in AI in schools is the European Driving License for Robots and Intelligent Systems (EDLRIS) . This European project followed the train-the-trainer approach to empower teachers to teach AI in classrooms. Within the project mini-curricula for Robotics and AI for a basic and advanced level had been developed. The intended duration of the courses based on the curricula are 5 days distributed over 3–5 weeks with alternating face-to-face and online units. Moreover, ready-to-use material for the different courses were developed comprising, for instance, background material, paper and pencil examples and coding examples. Furthermore, a certification system for teachers and students was developed. Certified teachers are entitled to give EDLRIS courses in their school as well as to use the provided material. Students who were educated in AI or Robotics by their teachers and had passed the certification can use this as an asset towards schools, universities or employers. This recognition, the openness of the Austrian computer science curriculum and the fact that EDLRIS courses are recognized as official training for the teachers, motivated numerous teachers to conduct a training.
An excellent example for informal education in AI K12 is the initiative Elements of AI (EoAI) . The initiative provides a MOOC that covers foundations of AI and targets a broad general audience. It provides the basic module Introduction of AI and the more advanced module Building AI. While the first fosters a basic understanding and covers definitions of AI, basic methods and implications, the second aims at a deeper methodical and mathematical understanding of approaches and a connection to coding. Although the first module also covers problem solving more generally the initiative has a strong focus on machine learning and big data. The units usually contain accessible descriptions using mainly text plus some figures and formulae and are concluded with tests in form of multiple choice questions, interactive items or online coding examples. Users are also able to obtain a certificate. Initially the initiative was a national attempt of Finland. Meanwhile due to its simplicity, promotion on the European Level and translations to other languages, EoAI gained huge attention. Over several hundreds of thousands of users enrolled to the course and a good fraction obtained a certificate. The initiative is also inclusive as can be seen by a good mixture of gender, age and nationalities of the users.
A different informal approach is the Technovation AI Family Challenge initiative . The central idea is to educate people in AI and to empower them to use this technology to solve societal challenges. The initiative aims at under-resourced communities and inclusion. Moreover, it fosters a strong participation of parents and guardians into the education process. More details about the initiatives can be found in . The methodology of teaching follows the idea of solving a real word problem using AI. Thus, there is a startup phase were concepts of AI and coding are taught. Later, the participants select a challenge from their local community like fixing problems with the water supply and work on a solution. A motivation factor is that there is a competition for the most innovative solution. Due to support of donating companies and foundations the initiatives partner up with local ambassadors and coaches which drives the initiatives locally. Following this strategy, over 20,000 people registered and around half of them completed the challenge.