The Act was initially designed to mitigate dangers from specific AI functions based on their risk level, but lawmakers pushed for its expansion to foundation models, which underpin general-purpose AI services. The Act also addresses AI-powered facial recognition surveillance systems, with a compromise allowing law enforcement to use them for serious crimes. However, rights groups have expressed concerns about the exemptions and other loopholes in the Act, including a lack of protection for AI systems used in migration and border control, and the option for developers to opt-out of having their systems classified as high risk.
Key takeaways:
- European Union negotiators have reached a deal on the world's first comprehensive artificial intelligence rules, which will provide legal oversight of AI technology.
- The Artificial Intelligence Act will still need to be voted on by the European Parliament early next year, but it is expected to pass as a formality.
- The Act includes provisions for generative AI systems like OpenAI's ChatGPT, requiring companies to provide technical documentation, comply with EU copyright law, and detail the content used for training.
- Despite calls for a full ban on public use of AI-powered face recognition surveillance systems, the Act will allow for exemptions for law enforcement to use them in cases of serious crimes.