The discussion on the implementation of artificial intelligence (AI) in education is increasingly focusing on ethical rather than technical aspects, and Europa can play a significant role in this. That was what experts Hanni Muukkonen and Keith Quille agreed on during the webinar on the European vision for Artificial Intelligence (AI) in Education. It is therefore no coincidence that after the summer the Commission is presenting a set of ethical guidelines that highlight exactly this aspect of the discussion. Brussels sees AI as a crucial determining factor in Europe's digital future, told Ioannis Gaviotis, who works at the Commission.

Europe must (and can) take the lead in the discussion on ethics in educational AI!

From left to right on the picture: Duuk Baten, Keith Quille, Hanni Muukkonen, Ioannis Gaviotis and Jurgen Rienks.

Responsible use of AI in European education

On Monday the 13th of June Neth-ER organised a webinar on the European vision for Artificial Intelligence (AI) in Education. The event took place as part of the Dutch "Month of AI in Education", organized by SIG AI in Education, the acceleration plan, the Dutch AI Coalition working group education, Kennisnet and SURF. Duuk Baten, Innovation Project Manager at SURF, welcomed everybody before handing over the word to director of Neth-ER and moderator of the session Jurgen Rienks. He engaged in a 50 minute round table talk with dr. Ioannis Gaviotis, policy officer at the Commission’s Directorate General for Communications Networks, Content and Technology; prof. dr. Hanni Muukkonen and dr. Keith Quille who – together with Duuk - are part of the European expert group on AI and Data in Education and Training.

The application of AI in education asks for more regulation

Europe is very interested in digital education, which is why there is the Digital Education Action Plan, told Gaviotis, explaining why the Commission focuses on AI in education. “We believe that with this [plan] we can bridge the digital divide and at the same time create more inclusive education.” One of the aims of the Plan is to enhance digital skills, that is why it is important to raise awareness on the risks and opportunities of AI and to create ethical guidelines. On the question why guidelines for AI in education are important, Gaviotis explained the details of the upcoming AI Act: “The AI Act follows a risk-based approach, in which it defines three categories of AI applications. Within this scope, education is categorized as a ‘high risk’ category for the use of AI. This means that there will be a regulatory framework covering the use of AI in education. Ethical guidelines aim at bringing forward the European set of values and helping educators by raising their awareness on this sensitive subject.

Shifting perspective in the AI in education discussion

‘What can Europe do for the responsible use of AI in education?’, asked Jurgen Rienks when turning to Keith and Hanni to talk about the guidelines. There is a great craving among people for what the application of AI in education means, Muukkonen responded. “That is why asking questions concerning ethical aspects of AI is an important starting point”. Quille agreed: “Having conversations like these today is amazing. Because it starts a cyclical process [of asking questions] and now companies are asking: how do I integrate ethics in my AI systems?”. This was also an important point for the outcome in the guidelines: “Teachers, practitioners, schools and companies should all be able to ask questions”. By providing educators with a toolbox to raise questions, the expert groups clearly hopes to raise the bar on transparency requirements for educational AI systems. This is how the conversation about AI in education is seeing a shift towards ethics, “where it used to be just technical skills […] I hope Europe will be one of the leaders in these types of conversations”, Quille added.  

United in difference: interdisciplinary approach needed

Although the guidelines will only be published in September, during the conversation some more tips of the hat were given about what to expect from the content. One of the components will consider an outline about what should be done when the use of AI in the classroom results in problems. “Teachers as users have a massive role in pinpointing problems with biased AI systems”, Muukkonen said. “So that is where teachers have an opportunity to help developers, by saying out loud where work is needed.” This also shows why a STEM background isn’t necessary to improve AI. Quille: “If we want to be serious about diversity and biases we need people thinking outside of their comfort zone, we [also] need social sciences and humanities on the forefront”.

Skill shortage

Luckily it was not hard to find a wide variety of experts for the expert group, as Europe has a vibrant community of researchers working in this field, Gaviotis told. But there is no doubt that Europe is short on people, Quille said, “especially in the field where the technical and ethics are combined”. Muukkonen sees that there is a lot of interest from young people for AI, but also notices they are being picked out by the more lucrative markets. But Europe also still works on improving the whole ecosystem of AI and education, in particular the Education Technology sector (or EdTech), Gaviotis stressed. The Commission is currently working on calls for this right now.

European ethics on AI in the world 

Looking outside of Europe, Quille concluded that Europe is at the forefront on integrating ethics in computing subjects. “The biggest challenge right now is integrating ethics in a sophisticated way in every computing course”, Quille told. “We are the first to not only integrate technicians in the development of AI, which is very positive”. Muukkonen also mentioned that there has been a ‘good slowing down’ of the technical developments, especially because the ethics questions first needed to be addressed. “Once we get those discussions out the way, we have a wonderful position for having really good ethical systems for AI in Europe”, Muukkonen concluded. At the closing Duuk Baten added that in the Netherland these discussion also have been held, and that the Value Compass created by SURF and Kennisnet helps organisations that have similar talks about the application of digital tools such as AI. Nonetheless all the speakers agreed that the conversation on the 13th was a good start, but still only a beginning.