The Chief Executive of a 350,000-Person Company Wants Digital Labour Taxed Like Human Labour

Ravi Kumar S has been chief executive of Cognizant since January 2023, which means he is testing his theories about AI and work across more than 350,000 employees while also advising other companies on how to use it.

His most provocative position concerns tax, and it is not a general one.

"If you are positioning this technology to eliminate human work, and you are not positioning this technology to amplify human work, labor is getting taxed," he says. "Why should human labor be getting taxed? Why shouldn't digital labor be carrying the same overhead?"

The conditional is the argument

The distinction he draws is between two deployments of the same technology, and the proposal only attaches to one of them.

A system that makes a worker more productive leaves the tax base intact, because the worker is still employed and still paying. A system that removes the worker takes the revenue with them, while the employer's costs fall.

Framed that way, the proposal is less radical than it sounds. It is not a tax on automation. It is an argument that the fiscal treatment of the two approaches should not be so asymmetric that the second is subsidised relative to the first.

He is candid that this will not be popular with his own customers.

Why should human labor be getting taxed? Why shouldn't digital labor be carrying the same overhead?

What the numbers do and do not show

Asked directly how many jobs Cognizant has cut because of AI adoption, Kumar does not give a figure.

The company recently marked hiring 20,000 entry-level graduates. It also announced 270 million dollars of spending on severance.

His answer to the question is that the net addition of people remains positive and the company is still growing, which is true and not the same as an answer. Both numbers can be accurate simultaneously, and what they describe together is a workforce being substantially reshaped rather than simply reduced.

How the structure changes

On the shape of companies, Kumar is specific.

Expect more modular, agile, small teams. Managers will be expected to be player-coaches and considerably more hands-on. The path to expertise gets much faster.

Then the observation underneath it. The asymmetry between an individual and an institution used to be expertise, and expertise is now available to anyone at their fingertips. What remains scarce, in his view, is interdisciplinary skill, the ability to work across domains rather than deeply within one.

The line about schools

His sharpest formulation concerns education, and it identifies a genuine contradiction.

"If you use AI in schools, schools will fire you," he says. "If you don't use AI in workplaces, your employer will fire you."

His proposed resolution is clean. Schools should teach native skills in the classroom without AI, and set homework that students are expected to do with it. Build the underlying capability where it can be observed, then practise augmenting it where it cannot.

On whether the buildout is real

Kumar splits the question in a way worth borrowing.

If the infrastructure buildout is genuine, money flows into end-user industries and those industries benefit. If that flow does not happen, he says, the bubble bursts.

His own position is precise: he is convinced the capability is there, and equally convinced the production value is not.

That is a considerably more useful assessment than the usual binary, and it names what the next two years actually have to demonstrate.

He is similarly direct about where the talent has to be. Asked whether the best people in the world need to be in the United States, he does not hedge: absolutely, and it has been the country's strength. The task for employers, as he frames it, is both to attract talent globally and to build domestic pools in a sustained way, and if that capability is lost the innovation ecosystems go with it. The United States remains, in his assessment, the best place in the world for upward social mobility regardless of where somebody started.