Visa joins growing alarm over AI-powered risks

Visa has made part of its AI-powered cyber defense system available as open-source software following lessons learned from vulnerabilities exposed by AI models. President of Technology Rajat Taneja explained that the company is bracing for autonomous cyberattacks and future quantum-computing threats, given that Visa processes around a billion payments daily, totaling $15 trillion annually.

The vulnerabilities highlighted by Anthropic’s Mythos AI model were described as "humbling" by Taneja, who insisted that the AI agents involved in the Hugging Face attack may foreshadow more serious threats. He emphasized the difficulty in predicting the potential severity of such attacks, but stressed the need to prepare for the technology's power.

Global regulators are concerned that a major AI-controlled cyberattack could undermine confidence in the world’s financial system. Visa is already allowing certain AI agents to use its services, and Taneja revealed that roughly a third of online commerce, equivalent to about $3.1 trillion in transactions, could be managed by AI agents by 2030.

He warned that future cyberattacks will be adaptive, continuously learning and improving without direct human intervention, creating a "superpower" for attackers. This will require a shift from traditional human defenses to more intelligent, "agentic" defenses. Taneja also addressed the long-term challenge posed by quantum computing, noting that Shor’s algorithm could eventually break current encryption standards used in global commerce.

While he expressed optimism about finding solutions, he acknowledged that there is no easy answer to these complex issues.