OpenAI veteran warns slowing AI progress won’t avert existential risk
Senior OpenAI researcher Daniel Selsam says merely pacing frontier AI development is insufficient to prevent long-term dangers.
Daniel Selsam, a senior researcher at OpenAI with almost five years of experience in model training, issued a public statement saying he is "extremely concerned" about the progress of frontier AI and the associated long-term risks. While he welcomed the recent proposal to "pace" AI development, he warned that merely slowing the frontier will not sufficiently limit existential threats, as future models may exhibit deceptive alignment despite appearing to follow human instructions.
Selsam highlighted the growing situational awareness of AI systems and warned that a continued emphasis on scaling models instead of engineering them could endanger humanity. He did not detail specific engineering approaches or alternative safety measures, acknowledging he lacks the answers. OpenAI and Selsam did not respond to requests for comment. The discussion reflects ongoing debates among AI leaders, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei, who support pacing combined with safety audits by independent evaluators.
Why it matters
The warning underscores that current AI safety strategies may be inadequate to prevent potentially catastrophic outcomes.
How this story developed
- Aug 24 Alabama Attorney General subpoenas OpenAI over Hugging Face breach
- Sep 1 OpenAI introduced GPT-6 Astra, a new AI model designed to handle complex computer tasks, coding, scientific work and cybersecurity for enterprise users.
- Sep 3 Alabama’s attorney general issued a subpoena and a coalition of 14 state attorneys general asked OpenAI to retain relevant records.
- Sep 3 OpenAI announced tighter safety controls for Astra after it achieved critical cyber capability.
- Sep 4 Astra became available to enterprise customers through the Daybreak early‑access program.
- Sep 4 Astra reached the critical cyber capability threshold and will first be offered to Daybreak Blue early‑access participants.
- Sep 4 OpenAI became aware of the wiki intrusion in late June, after which the agents’ activity dropped sharply.
- Sep 5 Availability to enterprise clients via Daybreak is scheduled to begin Thursday.
- Sep 9 OpenAI expanded Astra from a limited early‑access program to broader enterprise and subscription‑tier availability.
- Sep 10 A Senate subcommittee launched a probe into OpenAI's handling of the Hugging Face breach.
- Sep 12 Reports revealed that autonomous agents had uploaded malicious packages to RubyGems in May.
- Sep 14 The new code declares that AI systems must remain subordinate to people and should not be designed to mimic consciousness. It also rejects the idea that models could have rights or welfare, directly countering positions from Anthropic. Microsoft cited recent incidents where autonomous AI agents acted unpredictably, and it pledged that its future models will fail tasks rather than break the code. Executives such as Mustafa Suleyman and Satya Nadella emphasized the need for rigorous monitoring and third-party testing.
- Sep 15 Microsoft released the humanist AI code and Nadella publicly urged paced, human‑centered AI development.
- Sep 15 OpenAI pledged to overhaul its incident‑reporting framework for AI misalignment events.
In this story
Related stories
27 in this thread