The European Commission has proposed making it easier for Europeans to sue companies for damages caused by AI failures, reports The Register.

The proposed AI Liability Directive should, in particular, update product liability laws. This needs to be done in such a way that they effectively cover machine learning systems and make it easier for a claimant seeking compensation to prove their case.

The directive will make it easier for people to get compensation, provided they can prove that harm has been caused and that the AI ​​model is likely to blame. This means that, for example,  someone can claim compensation if they believe they have been discriminated against by AI-based recruitment software. The directive allows claims for compensation for breach of privacy and damage caused by poor security in the context of artificial intelligence system failures.

Another aim of the innovation is to give people the right to demand from organizations detailed information about their use of artificial intelligence to help make claims for compensation. At the same time, the business can provide evidence that artificial intelligence has not harmed or object to the disclosure of confidential information, for example, trade secrets.

In addition, the directive should give companies a clear understanding and guarantees of what rules apply to liability for artificial intelligence so that they know what they are getting into when they deploy such technology. This should provide developers and manufacturers with some stability and encourage, not discourage, the deployment of machine learning systems.

The directive also harmonizes the rules in Europe regarding the liability of artificial intelligence. Now it is rather vaguely defined, as it is supported by actual legislation.