In a recent commentary published by Project Syndicate, Rajan discussed a scenario circulated by equity research firm Citrini that warned AI could eliminate most white-collar jobs by 2028, with major consequences for the economy.
While such outcomes cannot be ruled out, Rajan said the prediction is “surely too pessimistic in some ways,” noting that new technologies historically take longer to spread across industries than expected.
“Outside a few sectors like software, frictions to adoption and sheer inertia will probably slow the pace of change,” Rajan wrote, citing examples such as automated telephone exchanges, which took decades to fully replace human operators.
Rajan later reiterated the point on LinkedIn, saying many AI forecasts overlook the role of public and political responses. “I think many do not take into account the possibility that society will have a voice on how it will play out,” he wrote.
Also read: The viral report that broke Wall Street has a scary prediction for India too
In his analysis, Rajan outlined multiple possible paths for the AI economy.
One scenario involves a small number of dominant AI platforms — such as systems developed by companies like Anthropic or Meta Platforms — gaining a strong technological lead and charging high prices to businesses that rely on them.
Such dominance could allow AI firms to earn large profits while enabling companies across industries to automate cognitive tasks and reduce white-collar staff.
Displaced workers could then move into service jobs such as retail or hospitality, increasing competition and pushing down wages in those sectors.
Rajan also described a different outcome in which multiple AI systems, including tools like ChatGPT and Gemini, offer similar capabilities.
In such a market, competition could keep AI prices low, allowing productivity gains to spread across the economy rather than being concentrated among a few firms.
Also read: Who is Alap Shah? The co-author of the viral AI crisis report roiling global markets
Lower prices for goods and services could partly offset job losses, while stronger demand could support employment in other sectors, Rajan said.
Governments could also influence which outcome emerges through policy decisions, including competition rules, regulation of AI pricing or taxation of excess profits from dominant platforms.
“If the AI-induced pain is indeed widespread, the political impetus for intervention will remain strong,” Rajan wrote.
Separate comments by Ethan Mollick, a professor at the Wharton School who studies artificial intelligence and entrepreneurship, point to another factor that could slow job disruption: the cost of computing power required to run advanced AI systems.
In a LinkedIn post, Mollick said many AI tasks, particularly systems that carry out complex workflows, require large amounts of computing capacity.
“That makes AI expensive,” he wrote, noting that engineers using advanced AI tools sometimes spend thousands of dollars a day on computing tokens.
Because of these costs, companies are likely to focus AI deployment on high-value tasks such as software development rather than replacing workers across a wide range of occupations.
“There will not be enough compute for many years to automate many human jobs, even assuming AI can do that work,” Mollick said.
Rajan said the most likely outcome could be a slower transition in which workers have time to adapt and learn how to use AI tools rather than being immediately displaced.
At the same time, competition among AI providers could prevent excessive concentration of profits and allow the economic benefits of the technology to spread more widely.
But he cautioned that governments and businesses should prepare for different possibilities as the technology evolves.
“Now is the time to map out the possible scenarios and start preparing for them,” Rajan wrote.
Also read: Anand Mahindra on rise of AI: ‘It won’t eliminate service providers, but…’
