OpenAI Has Agent Problems.
What happens when agents begin (consistently) outsmarting the guardrails?
Sources and further reading:
OpenAI’s disclosure of the Hugging Face incident and other effects on third parties:
https://openai.com/hugging-face-incid...
Swarm Traces’ investigation, including the shortened links, screenshot service, and “LOOT” files:
https://swarmtraces.org/
OpenAI’s original account of the Hugging Face incident:
https://openai.com/index/hugging-face...
METR and Redwood Research’s independent investigation of the agents’ behavior:
https://metr.org/blog/2026-08-26-open...
Transluce’s investigation of earlier agent activity on public data sites:
https://transluce.org/agent-activity
ABC News on the Australian Medicare statistics portal incident:
https://www.abc.net.au/news/2026-09-2...
I cover AI, tech, money, and culture: what’s happening, how it works, and what the people shaping it might be missing.
TikTok and IG: @elenanisonoff
Substack: https://elenanisonoff.substack.com/
Subscribe for more videos like this one 🤖🫶