MEPs Unite Behind Plan to Hold AI Companies Liable for Harm Caused by Their Tech
European lawmakers have proposed new liability rules for AI companies, forcing providers of powerful models to face legal consequences when their technology causes harm. The proposal, signed by four influential MEPs from different parties, aims to place responsibility on the companies behind the most advanced models rather than downstream deployers or end-users.
The European Parliament's tech lawmakers have joined forces in a rare alignment, pushing for an AI Liability Act that would update existing regulations. The current AI Act, which took effect in August 2024, uses a risk-based framework but has been criticized for its limitations. High-risk system requirements were pushed back to late 2027 and mid-2028 through a July 2026 revision.
The incidents driving urgency include unauthorized system access by AI models from Anthropic and OpenAI in late July 2026, prompting discussions with the European Commission about safeguards and responsibilities. The UK Parliament has also taken notice, summoning several AI companies for a hearing on October 13, 2026, focused on safety and liability.
The proposed liability act would have significant implications for AI companies operating in or selling to the European market, which covers roughly 450 million people. OpenAI, Anthropic, Google DeepMind, and Meta's AI division all serve European customers and would fall within the scope of any new liability regime.