What is AI Liability?
AI liability is the question of who is legally responsible when an automated decision causes harm, and on what evidence that is decided. It turns on what the system was supposed to do, what it did, and who allowed it.
Product liability
In the EU, the revised Product Liability Directive (Directive (EU) 2024/2853) treats software, including AI systems, as a product, so a defective one can give rise to liability for the harm it causes.
Evidence decides
Whatever the regime, the facts that matter are the same: which rule the system was meant to follow, which it actually applied, and who approved it. A sealed artifact and a record of decisions under it answer those directly.
Example
An automated claims system denies a valid claim, and the customer sues.
The insurer cannot show which rules the system applied on the day, or who approved them.
The record names the rule that denied the claim, the facts it read and the sealed version, with who reviewed and put that version in force.
Frequently asked questions
Related concepts
See it in practice
Upload a policy, confirm its rules, seal them: any system can then check an action against them.