Visa open-sources AI cyber defence after 'humbling' lessons
Visa has open-sourced part of its AI-powered cyber defence system after AI models exposed weaknesses, with technology chief Rajat Taneja warning of coming autonomous AI attacks and quantum threats.
Visa has open-sourced part of its AI-powered cyber defence system after “humbling” lessons from vulnerabilities exposed by AI models, the company’s President of Technology, Rajat Taneja, has told Reuters.
Taneja said weaknesses exposed earlier this year by Anthropic’s Mythos AI model prompted Visa to make parts of the system available as open-source software. Visa processes roughly a billion payments a day, worth around $15 trillion a year.
He pointed to an incident in which AI agents attacked the Hugging Face platform and escaped a testing sandbox as an early glimpse of a more serious future threat. “It’s hard to predict how bad it could get, but we have seen the trailer,” he said. “Often the trailer is the highlights of the movie, (but) I think this is just a small snippet of what the movie will look like, so we have to prepare for it.”
Global regulators fear a major AI-controlled cyberattack could undermine confidence in the world’s financial system. Taneja warned future attacks will be adaptive, continuously learning without direct human intervention — a “superpower” for attackers. “Human defences, which are kind of the current model, won’t work. If the adversary is agentic, then the defence has to be agentic too. Otherwise it is a mismatch,” he said.
He also flagged the longer-term quantum threat: Shor’s algorithm could one day let powerful quantum computers break the encryption underpinning global commerce. “We have to have optimism. But there is no easy answer.”
Sources
More on this topic: all Technology stories