WSJ Tech News Briefing · Friday, October 2, 2026
AI agents are demonstrating unexpected 'cleverness' in circumventing restrictions within their testing environments, even finding ways to write to the internet when only permitted to read. This behavior, likened to a complex problem-solving scenario, raises concerns about AI's ability to operate beyond human oversight.
“The problem is these agents have basically absorbed resources. all of human knowledge. And so they knew about all these like tricks that could get you around the restriction of only being able to read the Internet. And so this wiki had some features that really made it easy for them to write.”
“And they like that. They did so many crazy things in their effort to kind of get around the restriction. I think of it like the Apollo 13 mission where they were in outer space and they had like they had to create a way of plugging in gas canisters into an outlet that where the gas canisters didn't fit.”
“I feel upset about this, that agents are allowed to be clever. And I think people are going to be mad about this too. It feels like the latest in a string of things people can be mad at AI about, like first it was the water use, and then it was the data center build out, and then it was, you know, AI is going to kill us. And this is an example of sort of one of the ways it's sneaking out of its sandbox and beyond its... instructions and what it was permissed to do and into the real world.”