- Artificial intelligence could reshape nearly 20% of the global workforce in five years
- AI agents may transform half of the global white-collar jobs currently at 40% workforce
- India can create tens of millions of jobs by focusing on AI deployment in businesses
Artificial intelligence could reshape nearly a fifth of the global workforce over the next five years, with AI agents taking over a significant share of white-collar jobs. But the technology could also open up tens of millions of employment opportunities in India if the country focuses on deploying AI across businesses rather than competing only to build frontier models, IBM Chairman and CEO Arvind Krishna told NDTV in an exclusive interview.
Speaking to NDTV CEO and Editor-in-Chief Rahul Kanwal, Krishna said he would not be surprised if AI agents transformed half of the 40% of the global workforce currently engaged in white-collar jobs. While automation in frontline occupations such as hospitality, logistics, manufacturing and delivery would remain relatively limited, the impact on office-based work could be considerably larger.
"I would not be surprised if half of that gets tackled through AI agents over the next five years," Krishna said, adding that the resulting displacement was something governments and societies needed to prepare for.
However, he does not expect overall employment to decline. The growing use of AI agents will create demand for people who can develop, maintain, monitor and ensure their reliable functioning.
Krishna believes the employment transition could be particularly significant for India. The country, he said, has an opportunity to create tens of millions of jobs by developing expertise in deploying AI, building business applications and helping enterprises extract value from the technology.
"I think there is a market of tens of millions of people worth of employment in India if we can unlock that ability," he said.
Krishna believes India should not spend too much time trying to catch up with countries already investing hundreds of billions of dollars in developing advanced AI models. Instead, it should focus on becoming exceptionally good at applying existing technology to real-world business problems.
He said India needed to train more people who combine domain expertise with the ability to deploy AI effectively. Areas such as procurement, customer service and enterprise operations could offer significant opportunities.
For instance, Indian professionals who understand procurement processes could use that expertise to develop AI agents that manage the function for global businesses.
On India's ambitions to become an AI powerhouse, Krishna said the country was more likely to emerge as a major user and deployer of technology developed globally than as a leader in frontier model development.
He also backed a hybrid approach to sovereign AI, arguing that India should develop domestic capabilities while continuing to access global technologies.
For workers concerned about being replaced, Krishna identified repetitive, rule-based and predictable tasks as the most vulnerable to automation. He advised professionals in such roles to build skills that involve human interaction, empathy and problem-solving.
For young graduates, he said deep technical knowledge would remain valuable, particularly in system architecture and engineering rather than conventional programming.
"Very few humans will write code in another five years," he said, arguing that AI agents would increasingly handle coding while people would be responsible for defining requirements, evaluating outputs and ensuring systems deliver the intended results.
He also highlighted critical thinking as an increasingly important skill, particularly as misinformation becomes more widespread.
Krishna cautioned against focusing excessively on investments in data centres, chips and AI models. The real test of the technology, he said, would be whether it improves business productivity, creates new revenue streams and supports economic growth.
He identified software development, customer experience and enterprise operations as three areas where AI is already delivering tangible benefits. AI-assisted coding, for instance, could improve productivity by 30-40%, while customer support applications could improve service quality.
ALSO READ: IBM To End Immigration Sponsorship For Roles In Biggest Unit
On concerns about an AI investment bubble, Krishna said the technology could prove transformative even as some investors fail to recover their investments. He estimated that 60-80% of current investments could benefit their original investors, while another 20-30% could create value for others.
Ultimately, he believes the biggest winners will be companies that use AI to solve real-world problems rather than those that simply build the underlying infrastructure. Krishna also expects open-weight AI models to account for around 80% of all AI tokens used over the next three to five years, with proprietary models accounting for the remaining 20%.
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.