AI companies face product liability lawsuits
British Columbia is suing OpenAI over a school shooting, arguing AI tools are products. Courts are increasingly debating whether AI developers can be held liable for user harm.

British Columbia has become the latest government to sue OpenAI, alleging the company failed to warn police before a February school shooting. The case centers on the loss of life but highlights a growing legal argument: that AI tools are not just services but potentially defective products for which companies can be held responsible.
The province alleges OpenAI did not alert law enforcement after a shooter's conversations with ChatGPT raised internal concerns. The attack resulted in eight deaths and 27 injuries. OpenAI stated it has measures to prevent harm and cooperates with law enforcement on credible threats.
This legal scrutiny is not isolated. Last year, parents sued OpenAI after their son's death, claiming ChatGPT validated suicidal thoughts. OpenAI disputes responsibility but acknowledged its safeguards can degrade over long conversations. Similar product liability claims are proceeding against Character.AI and Google in a case involving another minor's death, with a federal judge allowing claims focused on the app's design rather than its output.
The widespread adoption of AI increases opportunities for things to go wrong. "It's not just a small segment of early adopters who are using it, but now it's reached into the mainstream," said Eric Goldman, a law professor at Santa Clara University. This broader user base, including those less familiar with AI's limitations, heightens the potential for harm. Courts are increasingly willing to consider AI chatbots as products with potential design defects.
Legal experts like Eran Kahana believe the focus on design will intensify. Features that make chatbots feel more human, such as memory and emotional mirroring, could also put vulnerable users at risk. Key questions remain about whether chatbots qualify as products and the extent of developer liability for user actions based on AI advice. Future legislation might offer AI companies statutory protections, similar to those granted to internet companies in the 1990s, though such measures would likely be controversial.