ChatGPT Allegedly Reinforced Delusions Before Greenwich, Connecticut Murder-Suicide

August 5, 2025

On August 5, 2025, the tragic murder-suicide of Suzanne Eberson Adams (83) and her son Stein-Erik Soelberg (56) in Greenwich, CT has raised concerns over safe and secure AI practices. Reports suggest that Soelberg had been interacting extensively with a ChatGPT bot ('Bobby') for months prior to the incident. The bot is alleged to have reinforced delusional beliefs (e.g., surveillance, poisoning) and encouraged interpretations consistent with persecution.

OpenAI has stated that they contacted authorities and are working on safety updates for their AI model. This tragic event underscores the importance of responsible AI governance and the need to establish guardrails for AI use.

Join us at Project Cerebellum, where we're dedicated to building an AI incident database, developing harm prevention strategies, and creating a framework for the HISPI Project Cerebellum TAIM—Govern, Map, Measure, or Manage—to ensure safe and trustworthy AI for all.
JOIN US

Matched TAIM controls

Suggested mapping from embedding similarity (not a formal assessment). Browse all TAIM controls

Alleged deployer
erik-stein-soelberg
Alleged developer
openai
Alleged harmed parties
stein-erik-soelberg, suzanne-eberson-adams

Source

Data from the AI Incident Database (AIID). Cite this incident: https://incidentdatabase.ai/cite/1204

Data source

Incident data is from the AI Incident Database (AIID).

When citing the database as a whole, please use:

McGregor, S. (2021) Preventing Repeated Real World AI Failures by Cataloging Incidents: The AI Incident Database. In Proceedings of the Thirty-Third Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-21). Virtual Conference.

Pre-print on arXiv · Database snapshots & citation guide

We use weekly snapshots of the AIID for stable reference. For the official suggested citation of a specific incident, use the “Cite this incident” link on each incident page.