Brussels: The European Commission is considering delaying parts of the EU's Artificial Intelligence Act following pressure from businesses and the Trump administration, as reported by various media outlets. The Commission has confirmed that discussions are ongoing about possible delays, following a Financial Times report indicating that Brussels is assessing potential changes aimed at easing requirements for major tech companies.
According to Anadolu Agency, the AI Act, which is the world's first comprehensive legal framework for governing artificial intelligence, was enacted in August 2024. However, many of its provisions have yet to be implemented. The act's requirements for companies developing high-risk AI systems, which could pose significant risks to health, safety, or fundamental rights, are slated to come into effect starting in August 2026 or possibly later.
The Financial Times highlighted that the Commission is contemplating a one-year grace period for companies that might breach the strictest AI rules. An internal document referenced by the Financial Times suggests that generative AI providers with existing market products could receive a year's reprieve to adapt their practices without market disruption.
Furthermore, the report indicates that the Commission is considering delaying fines for non-compliance with new AI transparency rules until August 2027, allowing AI system providers and deployers more time for adaptation. MLex news organization has also reported that Brussels is exploring more flexible, less prescriptive monitoring requirements for developers of high-risk AI systems.
These proposals are subject to change before their anticipated release on November 19, after which they will require approval from EU member states and the European Parliament, as noted by the Guardian. The Commission is facing pressure from the Trump administration to relax tech regulations, with the US president cautioning about potential tariffs on nations whose regulations might harm or discriminate against American technology.
Meta, the parent company of Facebook and Instagram, has declined to sign the Commission's code of practice for general-purpose AI models. Joel Kaplan from Meta criticized the code, stating that it creates legal uncertainties and exceeds the scope of the AI Act. Additionally, numerous European firms have advocated for a two-year postponement of the act, arguing it is necessary for reasonable implementation and further simplification.
In an open letter, CEOs from 46 companies, including Airbus, Lufthansa, and Mercedes-Benz, urged for a delay, asserting it would demonstrate Europe's commitment to simplification and competitiveness.