The European Parliament voted overwhelmingly today to adopt the AI Liability Directive, establishing a legal framework that holds developers and deployers of artificial intelligence systems accountable for damages caused by their products. The directive creates a presumption of causality when an AI system causes harm, shifting the burden of proof from victims to AI providers.
Under the new rules, companies deploying high-risk AI systems in healthcare, transportation, financial services, and law enforcement must maintain insurance coverage and provide affected individuals with access to technical documentation when filing claims. The directive complements the EU AI Act passed in 2024 by establishing the liability consequences for non-compliance with safety requirements.
Tech industry groups warned that the directive could slow AI innovation in Europe and drive companies to deploy new technologies in less regulated markets first. Consumer advocacy organizations countered that clear liability rules actually encourage responsible innovation by ensuring that companies invest in safety and testing. The directive must be transposed into national law by EU member states within 24 months, with full enforcement expected by mid-2028.
