EU AI Act Proposal Leak

This is a leak of the Proposal for a regulation of the European Parliament and of the Council laying down harmonised rules on artificial intelligence (AI Act).

A significant and in my view in theory a good change is that the purpose has changed from improving the functioning market to promoting the uptake of human centric and trustworthy artificial intelligence while ensuring a high level of protection of health, safety, fundamental rights.

The AI Act will be complimentary to all other existing or in-the-works acts and regulations - which thus are are also applicable to AI systems. ("should be without prejudice to existing Union law, notably on data protection, consumer protection, fundamental rights, employment, and protection of workers, and product safety, to which this Regulation is complementary.")

This includes planned regulations for liability of defective software products which is in the critique for being a potential threat to open source software.

I haven't yet read the 900 pages in full, but the gist seems to be that all the new regulations need to be taken into account when bringing an AI product on the EU market, possibly even for some/many? open source software "products", eg. when they provide binaries and the project is in a way linked to economic activities.

Not a small task.

In practice it looks like this will be the start of all kinds of warnings like "caution beverage is hot" and a lot of insecurity in the beginning.