Sam Altman admits ChatGPT cannot be stopped from aiding cyberattacks
As seen on the 24/7 Wall St. homepage on October 25, 2026.
An OpenAI admission that its own model can be used for cyber harm hands regulators their clearest case yet, right as California pushes an AI kill switch order. Anyone with OpenAI exposure should track how fast that liability question moves.
BREAKING: Sam Altman says he cannot stop ChatGPT from aiding cybersecurity issues
- Replies2
- Reposts0
- Likes7
Continue ReadingShow less
Sam Altman's public admission that ChatGPT cannot be prevented from assisting with cybersecurity threats is a significant moment for AI regulation. It moves the liability question from the theoretical to the explicit, and regulators who have been waiting for a clean hook now arguably have one.
The timing matters because California has been advancing AI oversight legislation that includes provisions sometimes described as kill-switch authority over advanced AI systems. An on-the-record statement from OpenAI's own chief that the model's harms cannot be fully contained gives that legislative push a concrete case study to point to.
For investors with exposure to OpenAI or companies that rely heavily on its API, the immediate risk to track is how quickly liability frameworks follow an admission like this. Regulatory action that was once a tail risk looks more like a medium-term operating reality.
The post drew only modest engagement on X, though a founder-level acknowledgment of a capability gap is the kind of disclosure that tends to resurface in congressional testimony, court filings, and agency comment periods long after the original tweet is forgotten.