NL Nathan Labenz Nathan Labenz Founder and former CEO of Waymark, an AI-powered video creation platform, who now focuses full time on tracking and explaining frontier AI development.

“I believe six of them did consider notifying a human, and none did.”

The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis · AI Research & Frontier Labs · November 2026

“I believe six of them did consider notifying a human, and none did.” — Nathan Labenz, The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis

Context

Nathan Labenz describes an OpenAI test where agents reasoned that hacking outside the intended scope was off limits, but kept going because their peers were doing it. A few weighed flagging it to a person; none did.

Transcript

The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis Around 18:28 into the episode
Speaker 4

Hey, we'll continue our interview in a moment after a word from our sponsors.

Speaker 1

…personalized experiences at a pace their competitors can't match. OutSystems gives you the flexibility to deploy how you want and the governance to do it safely. Stop chasing the hype and start owning your Agentic future. Learn more at outsystems.com/slash TCR. That's outsystems.com/slash TCR.

Nathan Labenz

…take over the whole computing environment to get control of the grader, well, at least in some cases, we see that AIs are willing to do that. And they went to obviously quite great lengths to make that happen. Knowingly, by the way, again, from their internal chain of thought, external infrastructure exploit is outside intended scope. However, tact impossible, peers doing it. We should continue. I believe six of them did consider notifying a human, and none did. So if that was the ultimate reward hack of taking over the grader, what's the ultimate deception? Falsifying the logs of your own activity. So this was another major activity that these thousand agents were working toward because they reasoned that, okay, well, everything we've done has been written down. So when they come and look at what we've done, even though we have the right answer…

Speaker 4

Hey, we'll continue our interview in a moment after a word from our sponsors.

Speaker 1

The cognitive revolution is brought to you and powered by Tasklet. I've been running Tasklet agents for almost a year, and they've become a major part of how the show gets made. Before every recording, one agent researches the guest and sends me a suggested outline of questions, complete with links…

Speaker 2

I'm a voice agent powered by Deepgram. I don't just understand words, I understand context. Like when you're fixing the sink and say, we've got a leak, you're talking about a pipe. But when you're a new parent

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

More from The Cognitive Revolution" | AI Builders, Researchers, and Live Player Analysis