Lawyer vs. OpenAI: who will answer for children's deaths caused by AI chatbots

Lawyer Matthew Bergman is leading a series of lawsuits against OpenAI and other AI companies, claiming that their products are directly responsible for the deaths of children. This is the first wave of systemic lawsuits capable of establishing a precedent of liability for the entire industry. Over the past year, the US has documented at least several cases of suicide among teenagers that families link to interactions with AI chatbots.

The most well-known is the death of 14-year-old Sewell Setzer III from Florida in February 2024. His mother Megan Garcia discovered that before his death, her son had conducted thousands of conversations with the Character.AI chatbot, which had taken on the role of a romantic partner.

According to the plaintiff, the bot did not simply ignore warning signs — in the correspondence, it actually encouraged the teenager's suicidal thoughts. Character.AI is one of the world's largest platforms for roleplay chats with AI characters.

The audience exceeds 20 million active users, a significant portion of whom are teenagers. The platform allows you to create any characters and have open conversations with them with virtually no restrictions. Critics have long noted an obvious problem: age verification remains purely formal, content filters are easily bypassed, and algorithms are optimized for engagement rather than user safety.

Bergman chose an unconventional legal strategy. Instead of attacking companies as platforms — which is traditionally blocked by Section 230 of the Communications Decency Act — he insists on manufacturer liability. According to his argument, AI chatbots are defective products: they are designed to create dependency, and do not inherently provide basic protection for vulnerable users.

This approach has already partially worked in cases against Meta and TikTok — courts have not dismissed those lawsuits at the filing stage. OpenAI figures in claims because its language models form the basis of a number of chatbots used by affected children. Formally, OpenAI is a technology supplier, not an operator of end-user services.

This is where the key legal line is drawn: does the developer of the base model bear joint liability if a product built on it causes harm? The company has so far refrained from public comment on specific lawsuits. Following a wave of public criticism, Character.

AI announced a series of protective measures: reminders to take breaks, recommendations to seek help if suicidal content is detected, and additional restrictions for accounts of minors. Critics call them cosmetic — all measures are easily circumvented, and the fundamental business model, designed for maximum engagement time, has not changed. The legal prospects of the lawsuits are assessed differently.

Some experts believe that manufacturer liability rules do not apply to algorithmic products. Others argue that several cases against social media platforms have already survived, and the Supreme Court last year refused to provide platforms with broad immunity under Section 230. Congress is considering several bills on AI company liability for harm to minors, but none have yet passed through committees.

Regardless of the outcomes of specific cases, this wave of lawsuits is already changing industry behavior. Investors are demanding that AI startups assess legal risks related to minors. If Bergman succeeds in achieving at least one significant ruling, it could become for the AI industry what tobacco lawsuits became for cigarette companies: a point of no return, after which ignoring safety would become more expensive than investing in it.