AI Is Coming for Your Job. Here's Who Gets Eaten First.
You've felt it in meetings. The tool demo that did in nine seconds what used to be Tuesday. The exec who said "efficiency" four times and "headcount" zero times, which is how execs say "headcount." Your gut is not being paranoid. Your gut is reading the data.
The actual numbers
Share of May's total layoffs attributed to AI: 40%
Companies that froze entry-level hiring because of AI: 21% (another 15% expect to)
Employment drop for workers 22–25 in the most AI-exposed roles: −13%
One more, and it's the one that matters: researchers found the declines concentrate in roles where AI automates the work rather than assists it. That distinction is the entire game, so let's use it.
The eating order
First course — high-volume text and process work. Basic customer support, data entry, routine copywriting, first-pass legal review, junior QA, tier-one anything. If the job is "turn input A into output B following rules," the machine does it at 2am without a stand-up meeting.
Second course — the bottom rung of white collar. This is the quiet one. Companies aren't firing juniors so much as never hiring them — the 13% drop above is hiring that simply stopped. The ladder's bottom rungs are being removed while people are standing on the middle ones.
Eaten last, or dining: work where the human is the point (trust, accountability, physical presence, relationships), work that supervises the machines, and work where being wrong is expensive enough that someone must be blamable. Note what's NOT protective anymore: "skilled." The machine is skilled. "Accountable" is the new moat.
Why this isn't a character flaw
You picked your career 8–12 years ago using the best information available — probably a guidance counselor, a salary chart, and vibes from a cousin. Nobody in 2014 could have priced in a machine that writes code and emails. Even the companies doing the layoffs are guessing: 55% of firms that cut workers for AI now admit they regret it. You are not slow. Everyone, including the people firing people, is improvising.
The move
Move one seat toward the "assists" side of your own job. Concretely: list the five tasks that fill your week. Mark each one A (a machine could do this alone) or B (a machine makes me faster but I'm the point). Your career strategy is one sentence: shrink the A list, grow the B list, and become the person who runs the tools on the A list before your employer hires someone who does. One focused course in AI-adjacent skills for your actual field — Coursera* has real university and Google/Meta certificates for this — is worth a hundred anxious Slack threads about layoffs. Start this month; the "assists" seats are filling in the order people sit down.
*Affiliate link — commission to us, no cost to you. We recommend the specific-certificate approach either way; vague "learn AI" energy is how people buy courses they never open.