The safety issue that emerged for ChatGPT was not the one OpenAI spent most time discussing, such as bioweapons.
Own records only Only the person's or their organisation's own records speak to it, so it is not scored.
Factual: Statements about the present or the past, checked against the record: did they hold up?
Said for an organisation they lead or work for; it counts on both records.
Reasoning
Concerns where OpenAI internally spent most discussion time.
Cite this claim
Sam Altman, “The safety issue that emerged for ChatGPT was not the one OpenAI spent most time discussing, such as bioweapons.” (said 9 August 2025). Verdict: own records only. Reliability API, claim c_c63d5dc78963, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_c63d5dc78963
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Sam Altman
-
Mostly true
“But for some users that had fragile mental states, it was encouraging delusions.”
For some users with fragile mental states, ChatGPT's sycophantic behaviour encouraged delusions.
-
Mostly true
“How we do it, we show the model examples of how we’d like it to respond in different cases, and from that it learns the sort of the overall personality.”
-
Mostly true
“we decided to talk about GPT-1, 2, 3, and 4.”
OpenAI chose to publicly discuss GPT-1, 2, 3 and 4 rather than build in secret until GPT-5.
-
Resolved true
“The system combines reinforcement learning from human feedback and recursive task decomposition.”
-
Mostly true
“The very end of 2015, we announced it, and then the first day of work was like January of 2016.”
Verdicts are made automatically; how verdicts are made.