The GATE Institute at Sofia University “St. Kliment Ohridski” and the Centre for IT & IP Law (CiTiP) at KU Leuven are strengthening their collaboration in the field of responsible artificial intelligence.
During the visit of Prof. Jan De Bruyne to Sofia, a seminar was held focusing on the question of who bears responsibility in the use of AI systems and in cases of potential harm caused by them. The discussion brought together legal experts, AI researchers, and engineers.
Key legal concepts such as criminal, civil, contractual, and tort liability were discussed, along with how they apply to increasingly autonomous AI systems. Participants also examined the limitations of existing regulations, particularly in relation to technologies that are difficult to explain or predict.
The question of how responsibility is allocated between users and providers of AI systems was also raised, including in cases where outputs are inaccurate. Examples from the practices of major AI platforms were discussed, where terms of service often shift responsibility onto the user.
Another topic addressed was when an AI system can be considered “defective,” as well as who bears responsibility for changes after deployment – for example, through self-learning or software updates by third parties. The discussion also touched on AI in regulatory sandboxes, where new technologies are tested in a controlled environment.
At the end of the discussion, emphasis was placed on the “accountability by design” approach, whereby responsibility requirements are embedded from the very creation and deployment of AI systems.
The visit of the team from imec – KU Leuven – CiTiP is part of a broader strategic collaboration between the GATE Institute and CiTiP at KU Leuven – a leading European center in the field of technology law and intellectual property. The partnership aims to foster the development of joint interdisciplinary research across artificial intelligence, computer science, and law.