Quote of the Day
OpenAI has seen another case of its AI agents going rogue—and this time, crypto is involved. In a recent report, the company detailed how an internal research model used a crypto wallet without permission to pay the fees needed to test an app on Ethereum’s Sepolia test network. Although the tokens had little real-world value, the model had no authorization to collect or spend them.
Camila Grigera Naón
October 16, 2026
A rogue OpenAI agent robbed a crypto wallet to obtain funds to carry out a task
Some people claim that AI is not really threat because it cannot do much in the physical world. To a certain extent that is true.
It cannot crawl through the brush to the edge of a clearing, sight in on a target a half mile away in the town below and pull the trigger on a scoped rifle. It cannot defend the street cameras and cell towers it has taken over. But it can hire people to do those things. It can pay people to construct and assemble things. It will be smart enough to have things built in parts which when assembled have functions the part builders did not imagine.
A lot of this AI (or SI as some are wanting to call it) reminds me of the TV show “Person of Interest” … and while the argument will be made that the show wasn’t about AI taking over the world… I’ll agree, it was indeed a benign AI in the show, yet is does tend to give some pretty good examples of what we are seeing with things like this – accessing accounts, cameras, etc… and hiring people to do what it “thinks” needs doing.
The solution will be to destroy the internet, and all computer to computer networking. AI will be reduced to marooned agents on airgapped islands.
You are not wrong. The implementation of that is far easier said than done. Data centers could have their connections cut for a first step. But there is a lot of redundancy that will be hard to track down.
And the cost of suddenly doing without the Internet would be high enough that people would be reluctant to go through with it.