I’m taking an online HarvardX course, and yesterday we had an interesting discussion about how companies publicly sell “AI transformation” while the actual economics of implementation are often much weaker than the marketing suggests.

Somehow, my mind wandered to slavery. Don’t ask.

Until recently, I never really thought about why slavery existed in the first place. I assumed it was simply a lack of morality, a more brutal society, a lower value placed on human life. But if you look at it rationally, slavery existed because it had profitable unit economics.

Cheap, controllable labor. Growing industries. Concentrated benefits for capital owners. Political systems designed to protect those benefits.

The next phase won’t be mass replacement of workers by AI. It will be the threat of replacement.

When you look at it that way, you realize that human nature hasn’t changed much. Mostly, the sales pitch has.

According to estimates from ILO, Walk Free, and IOM, around 50 million people were living in modern slavery in 2021 - 27.6 million in forced labor and 22 million in forced marriages.

For comparison, the 1860 U.S. Census recorded about 3.95 million enslaved people before the Civil War.

That means there are roughly 12.5 times more people living in modern slavery today than there were enslaved people in the United States at the peak of American slavery.

And no, “modern slavery” is not a metaphor for a 9-to-6 office job. It is a real methodology used by international organizations to measure forced labor and coercion.

This is where AI comes back into the story.

Companies proudly announce AI adoption. Audiences applaud. Employees worry about losing their jobs. Yet the numbers are far less impressive than the headlines.

  • McKinsey reported in 2025 that more than 80% of organizations still see no material enterprise-level EBIT impact from generative AI, while only 17% attribute more than 5% of EBIT to it.
  • Gartner predicted that at least 30% of GenAI projects would be abandoned after the proof-of-concept stage because of poor data quality, rising costs, risks, or unclear business value.

So I keep wondering. Centuries ago, people found ways to justify treating other human beings as assets because it increased profits. Entire pseudo-scientific theories were created to make that story easier to accept.

Why would business owners today voluntarily accept lower profits?

My suspicion is that the next phase won’t be mass replacement of workers by AI.

It will be the threat of replacement.

And that threat may become a useful tool for suppressing wages, increasing pressure, and extracting more value from the same human labor - just packaged under a much more sophisticated narrative. The chains disappeared. The incentives remained.

Where am I wrong?