The EU AI Act, which regulates artificial intelligence, is still being debated by European lawmakers. They are struggling to agree on how to regulate foundation models, and it is unlikely that the regulation will be passed before December. Spain, currently heading the EU, is pushing for more regular vetting for vulnerabilities and a tiered system of regulation based on the number of model users. Three trilogues have already been held, with a fourth expected this week. If no agreement is reached this month, another meeting will be scheduled for December, leading to concerns that decision-making on the law could be delayed until next year. One draft of the EU AI Act proposes that foundation model developers must assess risks, conduct testing, evaluate bias in training data, validate data, and publish technical documents before release. Some open-source companies have called for smaller companies to be considered in the discussion, arguing for a distinction between for-profit models and hobbyists/researchers. The EU’s AI Act has been seen as a potential example for other governments, including the US, in drafting regulations for generative AI. However, the EU has been slower in its progress compared to countries like China, which enacted its own rules in August.
>>Join our Facebook Group be part of community. <<