EU passes landmark AI regulations to ensure safety and transparency
-
EU countries have approved a comprehensive set of laws to regulate AI systems and ensure they are safe and respect rights. It bans high-risk models and puts obligations on developers.
-
The law exempts research-focused AI models, but could still encourage good practices around transparency and reporting biases.
-
Powerful general purpose models like GPT face transparency requirements and stricter regulations if deemed "high impact".
-
The act aims to incentivize open source AI by exempting certain models published under open licenses.
-
An AI Office will be created to oversee enforcement, but some question if it will have the resources to adequately scrutinize industry submissions.