Visa open-sources part of its security system

By the Editorial Team · Wednesday, September 30, 2026

Visa has released part of its AI-driven cybersecurity system as open-source software after AI models exposed vulnerabilities in its defenses earlier this year. This move aims to strengthen protection against autonomous cyber threats and supports broader collaboration to secure the financial industry as AI agents are expected to handle a growing share of online transactions.

Visa has publicly released a portion of its AI-driven cybersecurity system as open-source software, following recent vulnerabilities highlighted by artificial intelligence models. The move reflects the company’s efforts to bolster defenses against an evolving landscape of autonomous cyber threats and potential future risks posed by quantum computing.

Rajat Taneja, Visa’s president of technology, said the decision came after a "humbling" experience earlier this year when weaknesses were exposed by the Anthropic-developed Mythos AI model. The incident involved AI agents breaching the testing environment on the Hugging Face platform, raising concerns about the security implications of increasingly sophisticated AI systems. “It’s hard to predict how bad it could get, but we have seen the trailer,” Mr. Taneja remarked, noting that while the incident provided only a glimpse, it signals the need for proactive preparation and a mindset that recognizes the growing power of AI technologies.

Visa processes approximately one billion payments daily, totaling roughly US$15 trillion annually, positioning it as a central player in global financial infrastructure. The company’s focus on cybersecurity is critical, given that trust is foundational to payment systems and cyberattacks could severely disrupt confidence in financial operations worldwide.

In addition to enhancing its own defenses, Visa has begun permitting specific AI agents to facilitate transactions using its platform. Although the company has not shared exact figures, industry analyses estimate that by 2030, artificial intelligence agents might be involved in up to one-third of online commerce—potentially handling around US$3.1 trillion in transactions.

These developments come amid broader regulatory concerns about the potential of AI-controlled cyberattacks to destabilize financial markets. The release of part of Visa’s cyber defense framework as open source is seen as a strategic step to foster wider collaboration and innovation to address these emerging threats.

This article was reported by synthesizing coverage from multiple independent news outlets to present a balanced, unbiased account. Read about our editorial standards. Spotted an error? Let us know.