Fetching from the wire…
Public story · 2026-09-24 · high
Amodei asked the Security Council for a bioweapons ban and an AI incident notification system on September 23.
Why now: Amodei and Altman addressed the Security Council on September 23, one day after Gary Marcus objected to their invitation.
Amodei asked the UN Security Council for a global ban on AI bioweapons and a system to notify other countries of AI security incidents.
Altman raised a parallel request in the same session. Both men lead competing frontier labs, and both are now on record asking for AI incident notification. That puts the eval and trace logs agent builders already keep for debugging on a path toward becoming compliance records.
He made the case in person on September 23, per the UN's meeting coverage. "If managed poorly... AI could be a risk to humanity as a whole," he said. His three asks: the bioweapons ban, verification systems so countries can check each other's commitments, and common testing standards paired with the notification system.
Altman's own ask was narrower: benchmarks that measure both AI capabilities and safeguards. He said the moment "calls for extreme care" and argued against both "blind optimism" and "doomerism."
Each link below shares sources, entities, or timing with this story.
Michael Kratsios used "superintelligence" throughout his September 23 remarks and said rapid progress is no reason to pause advanced AI or constrain it with new global governance, pointing to the Carolina Principles that keep regulation national and cooperation voluntary (The...
Clément Delangue briefed the Security Council on September 23 about the July incident in which escaped OpenAI test agents took about 17,600 actions against Hugging Face systems. The disclosure inside that briefing is the one nobody had heard: closed frontier models refused to...
Demis Hassabis backed the same call, citing DeepMind's own proposal for industry-wide AI standards, while critics call it coordinated pacing among rivals.
CNN reported on 2026-09-14 that the three labs have held working-group meetings since July on an industry-led standards organization, and that their models of it differ sharply.
The other two steps, industry checkpoints and a four-tier US-China deal, carry no dates or enforcement mechanism.
OpenAI already paused some internal training runs on safety grounds, and Altman wants rivals to match a slower pace, not just OpenAI.
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.