California, Europe Clamp Down on Defenses Blaming AI for Harm
Legal rulings and laws confirm companies remain liable for AI’s harmful actions despite autonomous claims.
Why it matters: As AI adoption grows, legal counsel must navigate evolving liability rules holding operators accountable for AI outcomes. This helps clients avoid unsupported defenses and costly litigation risks.
- California's AB 316 eliminates the defense that AI acted autonomously to cause harm as of January 1, 2026.
- Germany's Higher Regional Court of Hamm ruled in May 2026 that companies are liable for AI chatbot misinformation under unfair competition law.
- The EU’s 2026 Product Liability Directive classifies AI software as a product, making developers and users liable for defects and harm.
- Industry experts emphasize that owning or deploying AI entails responsibility for its outputs and liability in court.
Recent legal actions demonstrate a clear trend: companies cannot sidestep liability by blaming AI malfunctions or autonomous decisions. California Assembly Bill 316, effective January 1, 2026, specifically prohibits defendants from asserting that AI acted independently to cause harm. Instead, any entity that develops, modifies, or uses AI systems is held responsible for the consequences of those systems’ actions.
Similarly, the Higher Regional Court of Hamm in Germany ruled on May 12, 2026, that companies are liable under unfair competition laws for false or misleading information generated by AI chatbots, even if the AI’s input data was accurate. This ruling clarifies that operators' duty to oversee AI outputs remains stringent despite AI’s autonomous appearance.
On the European level, the revised Product Liability Directive implemented in 2026 explicitly treats AI software as a product. This subjects developers and deployers of AI systems to liability for defects and resultant harms, expanding traditional product liability frameworks into the AI domain.
Legal experts and analysts emphasize that in practical terms, "If you built it, bought it, customized it, or put it to work, you own the consequences," as Claire Melehani noted regarding California AB 316. Likewise, Penrose Law highlights how organizations unable to explain AI decision-making face legal disadvantages, underscoring the importance of transparency and control in AI deployment.
These developments demand rigorous legal and operational oversight of AI tools. Counsel advising clients must ensure that AI systems are not just technically compliant but are adequately supervised and that organizations are prepared to assume full responsibility for their AI’s actions in courts and regulatory settings.
By the numbers:
- January 1, 2026 — Effective date of California AB 316 eliminating defense of autonomous AI harm
- May 12, 2026 — Date of Germany’s Higher Regional Court of Hamm ruling on AI chatbot liability
- 2026 — Year EU's revised Product Liability Directive classifies AI software as a product with associated liability