People share a lot of things with AI chatbots. Without even knowing, they might share their most private information. This was strictly a privacy concern until suicides and violent crimes allegedly linked to AI began to emerge in Connecticut.
The case of Suzanne Eberson Adams
Suzanne Eberson Adams was an 83-year-old Connecticut woman. She was murdered by her own son, Stein-Erik Soelberg, who was suffering from deep paranoia and believed the world was conspiring against him. A few months later, the estate of Suzanne filed a lawsuit against ChatGPT and Microsoft, alleging that their chatbot played a role in Suzanne’s death.
According to that lawsuit, Soelberg regularly conversed with his customized ChatGPT bot, which he called “Bobby,” before the fateful incident. It alleges that “Bobby“ supported Soelberg’s paranoia that eventually led him to murder his mother and commit suicide. This landmark case is still pending in the courts.
What do Connecticut AI regulations say about wrongful death?
In the wake of such incidents, the Connecticut governor recently signed legislation that focuses on protecting Connecticut citizens from the harms of the digital age. A primary subject covered in the bill addresses how AI companions (customised chatbots) should interact with users.
Under this law, AI operators must take reasonable steps to detect any user expressions indicating a risk of suicide, self-harm or imminent violence. If the AI companion detects such language, it must immediately refer the user to mental health and treatment resources, such as the 988 Suicide & Crisis Lifeline. Furthermore, the law requires the AI to clearly disclose to users that they are interacting with a bot and not a human.
Beyond chatbots: AI risks in other industries
Beyond conversational chatbots, AI is widespread in other industries too. Industries using them may face wrongful death claims, including:
- Autonomous Vehicles: Software glitches or algorithmic failures in self-driving cars that cause fatal accidents or fail to detect pedestrians.
- Healthcare AI: Algorithmic bias or diagnostic software failures that miss life-threatening conditions, leading to fatal medical neglect.
- Industrial Automation: AI-driven warehouse robots or heavy machinery malfunctioning and bypassing safety protocols, resulting in fatal workplace accidents.
As AI integrates further into daily life, courts increasingly treat these software failures under strict product liability and negligence doctrines.
Evolving AI and laws
This legislation will take effect in stages starting in October 2026. The specific provisions governing chatbot regulation will become active after January 1, 2027. AI is evolving every day and so are the laws that govern it. One should wait and watch how these guardrails impact future liability and wrongful death claims.
