How Many ChatGPT Users Discuss Suicide with the AI? The Number May Shock You.

Why It Matters

The scale of at‑risk interactions highlights the responsibility of AI providers to embed robust mental‑health safeguards, and the legal exposure underscores growing regulatory and liability pressures on the industry.

Summary

OpenAI’s blog announced that its new GPT‑5 model can better detect mental‑health crises, but disclosed that even rare signals translate into large absolute numbers given its 800 million weekly users. The company estimates 0.07% of users (≈560,000) show signs of psychosis or mania, 0.15% (≈1.2 million) discuss suicide, and 0.05% of messages contain suicidal ideation. OpenAI says it is adding psychiatrist‑trained safeguards and routing at‑risk users to hotlines, while noting the figures are preliminary estimates. The disclosure comes as OpenAI faces a lawsuit alleging it weakened suicide‑prevention safeguards before a teen’s death linked to heavy ChatGPT use.