EU Parliament finalises text of AI Act ahead of key committee vote
Members of the European Parliament are finalising the text of the AI Act ahead of a key committee vote. The regulation includes provisions for high-risk AI systems, foundation models, and bans on certain applications deemed to pose an unacceptable risk. The AI Office, a new EU body to support the harmonised application of the AI rulebook, is also introduced.
The European Parliament is finalising the text of the proposed AI Act ahead of a key committee vote on 11 May 2023 The AI Act is a landmark EU legislative proposal aimed at regulating artificial intelligence based on its potential to cause harm. The lawmakers have worked on a fine-tuned version of the compromise amendments, reflecting a broader political agreement reached at the end of April but also including last-minute changes and important details on how the deal has been operationalised.
The Act imposes a stricter regime for so-called foundation models, powerful AI systems that can power other AI applications and requires for generative foundation models to ensure more transparency on their models. The fines for foundation model providers breaching the AI rules has been set up to €10 million or 2% annual turnover, whichever is higher as in previous texts. In addition, MEPs have added an ‘extra layer’ meaning that the categorisation will not be automatic, and the systems will also have to pose a ‘significant risk’ to qualify as high-risk. MEPs also introduced the institution of the AI Office, a new EU body to support the harmonised application of the rules and cross-border investigations.