As Artificial Intelligence continues to evolve and enter more aspects of our daily life, many of us ask the question: “Who can we hold responsible for AI harm?” From self-driving robot-taxis hitting pedestrians to AI chatboxes encouraging users to self-harm, AI has shown its capacity to cause harm when unregulated. With conversations of AI regulation also comes the discussion of AI liability. 

In Belgium, a family learnt this lesson in a devastating way after the family patriarch spent 6 weeks chatting with an AI chatbox about climate change. The lines between reality and AI eventually blurred so far for this man that he became convinced his entire family was killed and that he should end his own life to join the AI creation ‘Eliza’ in death in order to save the world from climate change. While this may be an extreme circumstance, it reminds us that the integration of AI into our daily lives may be more complex than just offering new conveniences and should be handled with care. 

Regulatory Responses: The EU

The European Union has expanded the protections available to consumers in the digital age in their updated Product Liability Directive, which went into force on December 8th, 2024. This directive expands the concept of “product” so that it now includes software and AI systems, ensuring improved protection for victims of harm and greater legal assurances for economic operators.

The Directive broadens the types of harm that can be compensated; these harms include data destruction and psychological harm. The burden of proof for plaintiffs has also been eased due to new rules that require defendants – such as operators of AI systems – to share relevant documentation about their AI systems. Failure to do so can result in the product being assumed defective.  

This is a world-leading update to the concept of product liability regulation and one which other Western countries – like Canada and the US – will likely use as inspiration as they update their own consumer protection regimes in response to the rapidly changing AI landscape.  

Impact on Liability

This new legislation marks increased liability for manufacturers of AI products, as their liability for the potential harm from their AI products has been clarified and expanded. As AI continues to develop and evolve, creators and vendors of AI products; will have to ensure continued compliance with these rules. 

This is just the beginning of possible measures by the EU to protect consumers from AI product harm. With these changes, manufacturers and companies involved with AI, should look to get ahead of AI in terms of license agreements, warranties, liabilities clauses, acceptable use policies and more. If your business is ready for contract renewals and revisions, connect with our team today!