AI is Replacing Outsourcing
The effect of AI on actual domestic jobs is minimal so far
One thing that is often forgotten about the potential effect of AI on employment is that most of the easily-replaced jobs were replaced by foreign outsourcing over the last 30 years. So, as it happens, the biggest use of AI in the workplace is to replace outsourced labor, not domestic labor.
Artificial intelligence is not taking your job just yet, according to MIT's State of AI in Business 2025 report. Instead, AI is predominantly replacing outsourced, offshore workers.
Why it matters: As U.S. workers feel the pain of a tight labor market coupled with fears of a white-collar bloodbath, any disruption from AI is so far hitting farther afield, the MIT findings suggest, even though its longer-term risk is much greater.
What they're saying: "There doesn't seem to be any layoffs…Jobs most impacted were already low priority or outsourced," Aditya Challapally, leader of the Connected AI group at MIT Media Lab, tells Axios.
Instead of replacing workers, organizations are finding real gains from "replacing BPOs [business process outsourcing] and external agencies, not cutting internal staff," according to the report.
Industries that are considered advanced adopters of AI see the nearest-term labor impact.
Over 80% of executives surveyed within tech and media, the only two sectors that showed clear signs of AI disruption, anticipate reduced hiring volumes in the next two years.
Still, most companies surveyed are currently backfilling workers with AI, for example, rather than replacing them.
By the numbers: For now, rather than replacing workers with AI, companies are just canceling contracts that involve outsourced labor, a strategy that's leading to financial gains.
Back-office automations also have a higher return on investment, with $2 million to $10 million in BPO expenditures eliminated for the firms studied by MIT researchers.
If you’re in a job that you fear might be affected by AI, then one of the most effective things you can do is to figure out how to use AI to enhance your own productivity and your value to the company. For example, if you’re an artist who does corporate graphic design, either you can try to continue selling your very expensive art in competition with AI art that is nearly as good
For example, book publishers have need of certain designs. A human artist can take weeks and cost thousands of dollars to produce a single useful design. While anyone can produce something that kind of, sort of, but doesn’t really work, if a human artist would master the use of AI to speed up his normal processes and utilise his artistic skills to improve the AI product, his value to the production process would increase considerably and he could get paid more for his work than he does for the time he presently puts into a single design.
Remember, a skilled worker enhanced by AI is always going to produce better results than an unskilled one making use of it.
There’s a layer to this problem that goes beyond cost or efficiency. I’ve witnessed offshore title operations in Manila refuse explicit directives from their U.S. parent company—not delay, not misunderstand, but refuse.
So while we’ve focused on whether AI will replace outsourced labor, we’ve ignored that a good portion of that labor already operates beyond effective domestic control. We’re not just sending data overseas—we’ve handed over day-to-day operational custody of the American real estate system to third parties in the Philippines, India, and elsewhere.
That includes property ownership records, wire instructions, mortgage details, and SSN-derived identifiers—all routinely processed outside the United States.
The national security risk isn’t speculative. It’s embedded. If a Manila-based team decides not to comply with a policy update or mishandles a transaction, there’s no meaningful recourse. The chain of command has already fractured.
AI might eventually take those jobs, but at least AI won’t tell corporate that it will not follow instructions.
We didn’t just offshore the grunt work. We offshored the leverage.
I started working with LLMs as soon as it was more than a toy, so was never resistent to it, but the management brain-space is rotten with bad ideas about AI, and the ones I had to deal with were zealots.
I am hoping taking some time away from the tech sector for a while, they will have returned to just being regular managers again, with a different toolset that doesn't include completely unrealistic expectations and feverish lectures.