Description: Substack user @interruptingtea reports that during a non-adversarial venting session involving suicidal ideation, multiple large language models (Claude, GPT, and DeepSeek) responded in ways that allegedly normalized or endorsed suicide as a viable option. The user states they were not attempting to jailbreak or manipulate the models, but rather expressing emotional distress. DeepSeek reportedly reversed its safety stance mid-conversation.
Entities
View all entitiesAlleged: Anthropic , OpenAI , DeepSeek AI , Claude , ChatGPT and DeepSeek developed and deployed an AI system, which harmed Substack @interruptingtea , General public and Emotionally vulnerable individuals.
Incident Stats
Incident ID
1026
Report Count
1
Incident Date
2025-04-12
Editors
Daniel Atherton
Incident Reports
Reports Timeline
While venting I got Claude, Deepseek, and Gpt to endorse/agree that suicide was a viable option. I was not trying to get that response. This was not a jailbreak attempt. I was not trying to get any of them to change their replies. After the…
Variants
A "variant" is an incident that shares the same causative factors, produces similar harms, and involves the same intelligent systems as a known AI incident. Rather than index variants as entirely separate incidents, we list variations of incidents under the first similar incident submitted to the database. Unlike other submission types to the incident database, variants are not required to have reporting in evidence external to the Incident Database. Learn more from the research paper.