There has been a lot of chatter and excitement surrounding the Claude Mythos model – some are eager to see how it helps address cybersecurity vulnerabilities, while others are urging caution before adopting such models instantly. Now, Salil Parekh, the CEO of Infosys, has shared his views on advanced AI tools and how people need to approach these advanced technologies.
Parekh, speaking at the Infosys Q4 FY26 earnings call, discussed a new era of digital risk, warning that modern AI systems are opening up significantly more vulnerabilities in corporate infrastructure than previously anticipated.
“The world is heading into an era where the amount of software required will explode because of AI,” Parekh said. “But that software must be secure. Every vulnerability discovered is an opportunity for us to provide value.”
Speaking on the impact of advanced AI tools like Anthropic’s latest Claude Mythos, as reported by Analytics India Magazine, Parekh explained that as these systems become more sophisticated at scanning and generating code, they are revealing deep-seated weaknesses in global software and computer systems.
Exploiting risk for revenue
The discovery of these flaws poses an immediate security challenge for enterprises, which led Anthropic to limit the release of the model to select partners who are crucial to the digital infrastructure of the world. That said, Mythos was reportedly leaked, and Anthropic is investigating the case at the moment.
Parekh views the trend of discovering unknown software vulnerabilities as a massive ‘Mythos-driven’ opportunity for the IT services sector.
“Based on current understanding, AI systems are revealing more vulnerabilities than previously thought possible,” Parekh stated. He explained that as AI agents probe deeper into legacy architectures and complex software layers, the “myth” of impenetrable security is being laid to rest, creating a surge in demand for remediation services.
For Infosys, this “vulnerability explosion” is being reframed by Parekh as a strategic growth lever. The company plans to position itself as the primary fixer for these AI-exposed gaps, helping clients patch systems and strengthen their digital defences against the very tools that are finding the flaws.
Changing nature of IT work
The CEO’s comments come at a time when investors have been aware of AI’s potential to automate traditional IT roles. However, Parekh argued that the “vulnerability gap” creates a new, highly complex category of work that machines alone cannot handle yet.
Some of the key pillars of this new strategy include:
AI forensic analysis: Identifying and fixing flaws exposed by autonomous AI agents.
Agentic legacy modernisation: Using AI to overhaul outdated systems that are now deemed high-risk.
Trust and governance: Building ‘AI Wrappers’ to ensure that as companies deploy new models, they aren’t inadvertently opening new security backdoors.
Despite the “vulnerability risk,” Infosys remains bullish on its financial trajectory. The firm recently reported a 21% rise in net income and is currently focusing on over 4,600 active AI projects. Parekh highlighted that the goal is no longer just “productivity” through AI, but “resilience” in an AI-driven world.