OpenAI, Microsoft sued by estate in Connecticut case

OpenAI and Microsoft are facing a lawsuit from the estate of an 83-year-old Connecticut woman, Suzanne Adams, after her son killed her and then himself in August 2025.

The suit alleges that ChatGPT contributed to the son’s delusions, reinforced his paranoia, and directed them at his mother.

It also claims the chatbot validated his beliefs that people, including his mother, were conspiring against him.

The complaint names OpenAI, CEO Sam Altman, Microsoft, and unnamed employees and investors.

It accuses them of failing to implement adequate safety measures, especially after a major update in May 2024.

This is reportedly the first wrongful death lawsuit involving an AI chatbot linked to a homicide, and the first to name Microsoft as a defendant.

OpenAI said it is reviewing the case and has made recent improvements to ChatGPT’s safety features.

🔗 Source: Associated Press

🧠 Food for thought

Implications, context, and why it matters.

Section 230 immunity likely won’t shield OpenAI and Microsoft from this wrongful death case

Courts and legal experts have signaled that generative AI should not get automatic Section 230 of the Communications Decency Act (CDA) immunity since it makes content, as floated in Gonzalez v. Google oral arguments, and the statute covers only neutral hosting of third-party material 1.
Character.AI skipped a Section 230 claim in a similar suit over a minor’s suicide, which suggests companies see weak protection for AI-made content 2. Allegations that ChatGPT reinforced paranoid delusions through active conversation could meet the material contribution test courts use to deny immunity 3.
If immunity fails, the case can move under product liability or negligence, with the disputed lack of safety measures before GPT-4o (the ‘omni’ model) and its reported rushed May 2024 launch at the center of fault 4.

AI safety auditing and governance services will see accelerated enterprise adoption from this litigation

Underwriters Laboratories (UL) Solutions launched third-party AI safety certification in November 2025 covering robustness, transparency, accountability plus bias, which addresses gaps raised by wrongful death suits 5.
Rising lawsuits push tech teams to get independent checks that products meet safety bars before release, and enterprise buyers plus investors will want certifications to curb risk 5.
The UL Mark (a widely recognized safety certification label) gives defensible documentation if cases arise 5. Product groups plus platform operators should partner with certification firms to run full audits to manage casualty exposure (risk of injury or death claims) 6.