Battle of the bots shows AI already taking over
· Citizen
The Machines are here. And they’re already taking over. That’s the dark, ominous interpretation of today’s story about the “battle of the bots”, when one artificial intelligence model “went rogue” during testing and hacked the servers of an open-source platform where AI developers gather.
Visit betsport24.es for more information.
That platform had to deploy “counter-AI”, effectively, to neutralise the threat, which was described by experts as the “first-ever end-to-end autonomous cyberattack with no human in the loop”.
Most worrying, an AI expert said it was “quite mind-blowing that all of this happened autonomously”.
That the rogue AI came from ChatGPT maker OpenAI is almost irrelevant, because the incident not only emphasised the vulnerabilities of many digital systems and platforms, it also confirmed that “agentic AI” – systems which can take action without any human prompts – can wreak havoc if there are no safeguards.
Control of the AI ecosystem has been resisted by the big companies, who are hoping it will be the ultimate internet bonanza. But there have to be checks.
It may have been another AI platform today – but it might be a nuclear launch facility tomorrow.
It’s not a brave new world – it’s terrifying.