Will AI take your job? Wrong Question – Here’s who actually goes first
An idea I floated in an earlier post on the moral imperatives of AI about reskilling workers displaced by automation later surfaced in Indian government policy discussions. It also came up again, unprompted, in a recent keynote I gave for the PMI Chapter of Gujarat – will my job be replaced by AI?
Technological shift isn’t the only thing that renders people jobless. Policy changes and regulation do it too. The Covid shutdowns alone bankrupted businesses and erased jobs with no AI involved. Will AI cost jobs? Yes. Will it create jobs? Also, yes. The CxO question that actually matters is which roles inside your own organization are exposed first, and on what timeline, because the answer is not evenly distributed.
My prediction, and the one I’d put real money behind, is that the IT services sector gets hit hardest, reversing the pattern of every prior technology transition where it grew instead. Consider one of the roles large corporates typically outsource – Project and program management. An LLM trained on PMBOK and fine-tuned to your delivery lifecycle gives you a ready-made bench of junior PMs running small projects today, and increasingly sophisticated risk and coordination work tomorrow. Same with Product management roles. A model trained on your product line can draft PRDs and MRDs and spot cross-sell opportunities a human PM might miss. Software development becomes even more interesting. A model fine-tuned on your codebase doesn’t just accelerate development, it reviews the existing codebase, proposes fixes, and writes documentation and tests, the scope of work that currently occupies a full team.
The same pattern repeats across HR, sales and marketing, customer support, and legal. Call it the Exposure Order. The roles most exposed first are not the lowest-skilled, they’re the ones with the most standardized, well-documented inputs and predictable, often deterministic outputs. This is precisely why highly skilled, credentialed roles are more exposed than most executives assume.
That has a follow-on question worth sitting with. Does a two-year, six-figure master’s degree still make sense when a model trained on most of humanity’s available knowledge can serve as your co-pilot instead? If you’re building a workforce plan for the next 24 months, mapping your org chart against the Exposure Order, not against job titles, is where I’d start.