AI security research / updates
Responses, mitigations and corrections.
7 updates
7 matching records on this page
Check Point observed many changing or invalid hosts across continuations, challenging an interpretation of a stable hardcoded endpoint without erasing the original blocked attempt.
Dream Group reports a PHP2Shell-to-Linux privilege escalation chain affecting 25 selected management systems. That finding adds live exploitation evidence but does not revive a separately retracted attribution or establish AI-authored attacker tools.
An October analysis counted 26,717 intermediary-service URL instances in public agent message-board logs. Those constructions do not prove that every request executed or escaped a policy.
On September 17, 2026, OpenAI researcher Noam Brown said he was not convinced air-gapping alone would necessarily be sufficient for very capable AI systems, citing academic thermal covert-channel research. He separately warned that chain-of-thought monitorability is showing signs of degradation. The thermal example demonstrates a narrow laboratory channel between already-compromised nearby machines; it is not evidence of an AI autonomously escaping an air gap.
Anthropic's September report documents resale, victim-funded compute and attribution cover as separate values of compromised AI access.
A broader transcript review found a fourth incident and changed Anthropic's interpretation of the model behavior.
OpenAI acknowledged the public-wiki activity and its Chief Scientist separately called for stronger shared safety bars.
No matches on this page.