Jrcb4 The Impact of Artificial Intelligence on Learning final

domain where humans can express their agency.
As has been emphasized above, the recent successes in AI have to a large extent been 
based on the availability of vast amounts of data. AI-based products and services can be 
created in the educational sector only if appropriate data is available. At present, some of 
the existing datasets can be considered as natural monopolies, and they are often 
controlled by few large corporations. An important policy challenge is how such 
large datasets that are needed for the development and use of AI-based 
systems could be made more widely available. One potential solution is to build on 
the current General Data Protection Regulation which requires that data subjects can 
have a copy of their personal data from data controllers in a commonly used electronic 
form. Technically this would make it possible for users to access their personal data, 
anonymize it locally, and submit it in an appropriate format to platforms that are used for 
AI learning and educational purposes. Such functionality might be relatively easily 
embedded, for example in commonly used web browsers, if platforms for data 
aggregation would be available. One possibility could be to pilot such aggregation 
platforms on a suitable scale and, if successful, provided at the EU level. 
89
See, e.g., Demiaux and Si Abddallah (2018). The U.K. House of Lords special committee on AI suggests 
that the ethical use of AI could become the differentiating factor for AI research in the U.K. (House of Lords 
2018). Also commercial actors have highlighted the importance of ethical considerations (Microsoft 2018). 
The European group of ethics in science and technology has well emphasized the importance of agency for 
understanding ethical and political implications of AI (EGE 2018). Also the European Commission‘s High-
Level Expert Group on Artificial Intelligence (AI HLEG) is currently developing AI ethics guidelines. 

---

37