Published on December 7, 2023, 10:20 am
Thursday is a critical day for the AI Act negotiations in Brussels. The tech world eagerly awaits the outcome of the discussions, which began on Wednesday but did not reach a conclusion. However, there appears to be progress regarding generative AI systems like ChatGPT. Negotiators have reportedly reached a compromise on their control.
According to sources familiar with the talks, the discussions will now shift focus to the use of AI in biometric surveillance, an area that lawmakers want to ban. Reuters reports that some concessions may have been made by governments to allow for the use of AI in areas related to national security, defense, and military. These negotiations are expected to continue for several more hours on Thursday.
The AI Act is a groundbreaking attempt at regulating artificial intelligence globally and has been under development since April 2021. However, technological advancements, especially in generative AI represented by GenAI, have caused complications in its implementation.
Member states have also taken different positions on certain aspects of the act. Lawmakers have proposed requirements for developers to disclose information about how they train their models and indicate if copyrighted material is used or if content is generated by AI rather than humans. France and Germany, home to prominent European generative AI companies Mistral AI and Aleph Alpha respectively, oppose binding rules that could disadvantage these companies. Along with Italy, they advocate for self-regulation among developers based on a code of conduct.
If Thursday’s talks fail to yield definitive conclusions, concerns arise that the entire act could be postponed until after the European elections next year when a new Commission and Parliament take office. Given the rapid pace of technological developments like Google’s Gemini and ADM’s new super AI chip, regulators may need to rewrite the rules entirely by then.
While negotiations continue in Brussels, TNW’s senior editor Linnea provides comprehensive coverage of clean and climate tech as well as AI and technology politics. With a background in international relations, Linnea brings informed insights to her reporting.
In conclusion, the outcome of Thursday’s negotiations will determine how generative AI systems like ChatGPT are controlled under the AI Act. The discussions also focus on banning AI in biometric surveillance while allowing its use in national security and defense. Member states have differing views on regulation, with some advocating for self-regulation by developers. If conclusive agreements aren’t reached, the act may be delayed until after the European elections next year. TNW’s senior editor Linnea covers these developments and other crucial aspects of technology.