The European Commission published its long-awaited proposal for a Directive on Artificial Intelligence (AI) Liability. The draft directive aims to supplement the EU's AI law, still being negotiated by the EU institutions. At the same time, the European Commission also published its proposal to update the EU Product Liability Directive from 1985. If approved, the combined proposals will change the rules of liability for AI software and systems in the EU.

The draft of the Artificial Intelligence Liability Directive establishes rules applicable to non-contractual, fault-based, civil claims involving artificial intelligence systems. The proposal establishes rules that will regulate the preservation and disclosure of evidence in high-risk AI cases, as well as rules regarding the burden of proof and appropriate rebuttable presumptions.

In addition, the draft directive establishes circumstances in which a court may assume a relationship between (a) the fault of the provider or the user of any artificial intelligence system (whether "high risk" or not), and (b) the output produced by the artificial intelligence system or its failure to produce such output. For high-risk artificial intelligence systems, this assumption applies if the claimant has proved the supplier's or user's non-compliance with certain obligations according.

The Defective Products Directive gives courts the power to order suppliers or users of high-risk artificial intelligence systems to disclose (or retain) information about their systems to people who request that information to assert (or decide whether to assert) proceedings against the supplier or user. The EU's draft directive on defective product liability aims to hold manufacturers liable for certain damages caused by a defect in their "products" (which include both completed products and sub-components of products).

Click here to read the draft AI Liability Directive.

Click here to read the draft Directive on liability for defective products.