Infosys chair warns AI will create new legacy systems while cleaning up old ones

Infosys chairman Nandan Nilekani has issued a stark warning about the double-edged nature of AI adoption in enterprise IT, predicting that while artificial intelligence will finally enable organizations to clean up decades-old legacy systems, it will simultaneously create an entirely new generation of AI-generated technical debt.

Speaking at Infosys AI Day last week, Nilekani painted a picture of the current state of enterprise IT spending that many CIOs will recognize. "Many large companies are spending 60 to 80 percent of their IT spend on maintaining systems," he said, describing the massive drag that outdated infrastructure places on innovation budgets.

The Infosys chairman argued that AI represents the first real opportunity to break this cycle. "If you really want a firm to take advantage of AI, you have to fundamentally clean this up," Nilekani explained. "But the good news is, for the first time, because of AI, we have the tools now to do modernization... very quickly and in a much more economic way."

However, Nilekani's optimism about AI's cleanup capabilities is tempered by a sobering prediction about what comes next. "The very fact that you can generate stuff means you can generate slop," he warned. "In fact, five years from now, there'll be more AI legacy systems than any other legacy system – all the kind of stuff that will have been generated – and we'll have to clean that up as well."

This warning about AI-generated legacy systems reflects a growing concern in the tech industry about the quality and maintainability of AI-generated code. As tools like GitHub Copilot, Amazon CodeWhisperer, and other AI coding assistants become more prevalent, organizations may find themselves drowning in hastily generated but poorly architected software.

Nilekani also addressed the productivity paradox that many organizations are already experiencing with AI tools. He described a scenario where employees use AI to create the appearance of productivity without actually achieving meaningful results: "Let's say there are two guys and they are having a fight. One guy will draft an email that is one paragraph. He will give it to AI to make it into a ten-paragraph email because he wants to impress the other guy. The other guy will take the 10-paragraph email and summarize it back to one paragraph."

"So both have used AI, but what have we achieved? Nothing," Nilekani concluded, emphasizing the need for organizations to implement "usage guidelines, quality gates, and explainability rules" when deploying AI tools.

The Infosys chairman's comments come at a time when many organizations are grappling with how to balance AI adoption with maintaining code quality and developer productivity. His warning serves as a reminder that while AI can be a powerful tool for modernization, it requires careful governance and architectural oversight to avoid creating new problems while solving old ones.

Nilekani's prediction about AI legacy systems may prove prescient as organizations rush to adopt generative AI tools without fully considering the long-term maintenance implications. The challenge for IT leaders will be leveraging AI's modernization capabilities while establishing guardrails to prevent the creation of unsustainable technical debt.

As AI continues to transform software development practices, Nilekani's advice suggests that successful digital transformation requires not just adopting new tools, but also evolving organizational processes, quality standards, and architectural principles to match the capabilities and risks of AI-powered development.