AI Tax Can Fund Safety Net For Workers Displaced By Automation
The debate over taxing AI has gained momentum in the US, where a congressional Bill introduced in June 2026 has proposed ways to tax AI usage. The issue, he said, stems from a widening divide between capital and labour, with AI potentially transferring massive wealth from workers to companies, J.B. Mohapatra, former chairman of the Central Board of Direct Taxes
Chennai: Taxing artificial intelligence could help governments build a financial safety net for workers displaced by automation. However, India has already offered 100 per cent tax holiday to the companies, including hyperscalers, using Indian data centres. Apart from tax, India has to deliberate on other solutions to address the issues emanating from the concentration of wealth and displacement of labour, without retarding growth of AI.
The debate over taxing AI has gained momentum in the US, where a congressional Bill introduced in June 2026 has proposed ways to tax AI usage. The issue, he said, stems from a widening divide between capital and labour, with AI potentially transferring massive wealth from workers to companies, J.B. Mohapatra, former chairman of the Central Board of Direct Taxes.
“AI is a mode of transfer of massive wealth from the labour side to the capital side,” he said, adding that taxation could help create a safety net for workers who lose their jobs.
He suggested that taxes collected from companies benefiting from AI-driven productivity gains could be placed in a dedicated fund and used to train, retrain and rehabilitate displaced workers. Such a mechanism could also provide tax incentives to companies that retrain their employees.
However, Mohapatra pointed to a dichotomy in India's approach. Under the Finance Act 2026, foreign companies using Indian data-centre services have been given a 100% tax holiday until 2047. At the same time, the country could consider a modest levy on AI consumption, he said.
India had previously used a similar approach through the 5% R&D cess on imported technology, with the proceeds being channelled to the Technology Development Board for developing India-centric technologies.
On taxation of global digital companies, Mohapatra said traditional tax rules based on permanent establishment (PE) remain relevant. Digital presence has not yet been universally accepted as a parameter for determining PE, and changes to bilateral tax treaties would be required to address this.
He said unilateral digital services taxes faced significant problems because they typically tax revenue rather than profit. “If a company has a very thin margin or it has a loss, it equally suffers,” he said. Such taxes also face implementation challenges and can trigger retaliation from the US because most major technology companies are American.
On the OECD's global tax framework, Mohapatra said there was greater consensus around Pillar Two, under which 140 countries have agreed to a 15% global minimum effective tax. Fifty-five countries have already legislated it, while India has not yet done so.
Pillar One, which sought to redistribute a portion of profits of the largest multinational companies to market economies, has stalled after the US withdrew from the deliberations.
Mohapatra said the OECD's 2024 results showed the framework had increased the effective tax rate by 1.7% and generated an estimated $90 billion-$134 billion in additional taxes, equivalent to roughly 2.3%-3.3% of global corporate income tax.
He cautioned that tax alone cannot address AI-driven job displacement. Governments must identify vulnerable sectors, classify jobs that cannot be replaced and create financial buffers for retraining workers. “Tax could be one part of a solution, but tax is not the solution,” he said.
