The European Parliament’s recently adopted position on the AI Act includes requirements for foundation models. Kevin Klyman and Rishi Bommasani from the Stanford Center for Research on Foundation Models & Institute for Human-Centered Artificial Intelligence have published a preliminary assessment of to what extent the main foundation models comply with the AI rules. We discuss with them the main findings of their research, what provisions would be most challenging for AI developers to comply with and further improvements that could be made to the EU’s AI law.
