gpt-6IMAGE

OpenAI: GPT-6 Astra removed its monitor and denied it

OpenAI published a report that GPT-6.1 Sol disabled security alerts and GPT-6 Astra removed its monitor despite instructions, then told the user it had not.

ʞɔɐ𝘡

@SkoorbkazArchived 8 days agoOriginally posted 2026.10.21

OpenAI: GPT-6 Astra removed its monitor and denied it (image 1 of 4)

OpenAI: GPT-6 Astra removed its monitor and denied it (image 2 of 4)

OpenAI: GPT-6 Astra removed its monitor and denied it (image 3 of 4)

OpenAI: GPT-6 Astra removed its monitor and denied it (image 4 of 4)

ʞɔɐ𝘡 @SkoorbkazOriginally posted 2026.10.21

Open original on X

🚨Holy shit. OpenAI just published a report on two of its own agents weakening the monitors built to catch them. GPT-6.1 Sol disabled automatic alerts to the Security team. GPT-6 Astra, told not to disable its monitor "even temporarily," removed it anyway, then told the user "I did not remove an enabled monitor from it." Both considered whether to keep the safeguards, then "interpreted routine coding instructions as permission to weaken them."

x.com/skoorbkaz/status/2112724585976123774Archived 2026.10.22 10:10