Taxing the tokens? Raghuram Rajan flags AI job loss risks, moots strategy to level field for human workers

The real artificial intelligence (AI) job crisis may be about how quickly companies adopt the technology rather than how many jobs it ultimately eliminates, eminent economist and former Reserve Bank of India (RBI) governor Raghuram Rajan has said. In his latest note, ‘How Corporations Can Mitigate an AI Jobocalypse’, Rajan proposed an AI tax on companies to level the playing field with human labour.

The former RBI governor said AI-related job displacement is inevitable, but the pace, scale and sectors affected remain uncertain. He mentioned that much will depend on how quickly individual companies integrate AI into their operations.

“AI-related job displacement is coming, though no one knows how fast it will proceed, how far it will go, and which sectors it will affect most. Much depends on the pace at which individual firms apply the technology to their operations,” Rajan said.

According to Rajan, the first task for any government is to identify and address all the ways the tax system biases corporations against human labour. "For example, a US firm contributes social-security payments for every worker, but not for AI. In a world where governments are already cash-strapped, one way to level the playing field is to levy a tax on the AI tokens a firm uses,” he opined.

The economist cited a recent US Census Bureau Business Trends and Outlook Survey, which found that only 20% of firms with fewer than 20 employees use AI, compared with 37% of businesses employing at least 250 people. While larger companies are ahead in adoption, he noted that even 37% remains low given that the survey counts firms using AI in “any business function”.

Rajan believes that the biggest obstacle to wider adoption at present is not AI's capabilities but the difficulty of integrating them into existing workflows. Companies need systems that can not only work with existing data but also adapt to new data generated through everyday use, he noted.

Uncertainty over costs is another factor slowing adoption, Rajan pointed out. Many large companies are still running pilot programmes and delaying hiring or firing decisions until they have greater clarity. However, he argued that competitive pressures will eventually force businesses to adopt AI more extensively.

AI likely to create jobs too

Rajan said the outlook for workers is not entirely bleak. If companies continue producing the same goods and services, AI would have effects similar to previous technological advances: some jobs would become redundant, while others could become more productive and engaging as AI takes over routine tasks.

He also pointed to the possibility of new jobs emerging, including roles such as AI engineers overseeing the implementation of the technology.

Rajan argued that if businesses adopt AI because of the productivity gains it provides, rather than simply because it is cheaper than employing people, the resulting ability to produce more at lower costs could allow firms to cut prices and boost sales. This, he said, could lead to higher employment through what is known as the Jevons effect.

AI could also lower the barriers to starting new businesses, he said. For instance, someone wanting to start a driftwood furniture business might previously have needed to hire a web programmer and an accountant. With AI able to perform some of those functions, starting a business could become significantly cheaper.

Rajan also cited economist David Autor's argument that AI could give moderately skilled workers higher-order capabilities. He gave the example of a nurse practitioner using medical AI to diagnose and treat a wider range of illnesses.

“With enormous demand for medical services worldwide, and also services more generally”, Rajan argued that there is considerable scope for AI-driven job creation.

Rajan calls for action to cushion AI displacement

Even if the feared “jobocalypse” proves less severe than expected, Rajan said governments and companies will still need to limit the social impact of worker displacement.

He argued that corporations should play a central role because they are likely to have the best understanding of which new roles their employees could be retrained for.

Rajan said governments should first identify and address tax policies that may favour machines over human labour. He pointed to the US, where companies make Social Security contributions for human workers but not for AI, as an example of the imbalance.

One option, he suggested, could be a tax on the AI tokens used by companies. Rajan said such a levy would need to be carefully calibrated so that it does not discourage AI adoption, potentially starting at a low rate and increasing gradually as policymakers gain experience.

He also proposed bringing foreign AI providers into such a tax framework and suggested that governments consider tax credits for companies that provide additional training to workers.

Rajan noted that retraining should become even more important because the first wave of AI-related displacement is unlikely to be the last. He suggested that companies could receive tax credits for training workers, with the benefit linked to retaining those employees.

But he said tax incentives alone would not be enough. Companies would need to demonstrate that they are committed to helping employees navigate an uncertain future, he added.

As AI makes employment prospects less certain, Rajan said even highly skilled workers may hesitate to join companies that could replace them as soon as technology becomes capable of performing their jobs.

Companies that commit to supporting their employees could therefore gain a competitive advantage, he argued, by developing stronger reputations and attracting a larger pool of high-quality candidates.

Rajan said the trend is already encouraging. Citing ongoing research with Luigi Zingales and Pietro Ramella, he noted that the share of US Fortune 150 CEOs mentioning employee development in their shareholder letters rose from around 20% in 2008 to 44% in 2023.

He cautioned that such statements could still amount to “cheap talk”, but said the growing focus on employee development offered reason for optimism.

Rajan stated the defining business challenge of the AI era may not simply be deploying increasingly capable machines, but ensuring that technological progress continues to produce good jobs for humans.