On September 17, Indeed's Hiring Lab published an analysis of millions of US job postings with advertised salaries, covering 2021 through 2026. It compares pay in occupations most exposed to generative AI against the least exposed, using Indeed's own index of how much generative AI could change the way a job is done. One caveat belongs next to it, and the authors raise it themselves: part of the gain they measure comes from a shift toward fewer, more senior postings, and in their most conservative model the pay premium stops being statistically significant.
Even with that caveat, the headline number holds. Since 2021, advertised pay in the most AI-exposed occupations, among them software development, IT support, data and analytics, marketing and finance, has risen 46%, against 25% in the least-exposed occupations, such as nursing, caregiving, food service, cleaning and manufacturing. Business Today, which covered the study four days later alongside its own layoff data, summed it up as a paradox: the jobs AI touches most are also the ones paying the most, and the ones showing up most often in AI-linked layoff announcements.
Where the raise actually lands
The breakdown by seniority is the part that matters most to anyone hiring. Among AI-exposed occupations, the gap over less-exposed ones is 17 points at senior level, 12 points at mid-level and just 2 points at entry level. And within those same exposed occupations, the entry-level share of postings fell from 29% to 10% between 2021 and 2026, while the senior share rose from 22% to 47%.
What the second source adds
Business Today adds two things Indeed does not measure. First, layoff-tracker data from Challenger, Gray & Christmas: 116,175 AI-linked job cuts announced in the US through August this year, 22% of all cuts, including cases such as Block's planned reduction of 4,000 roles. Second, a warning from the International Labour Organization that shrinking entry-level postings risk eroding job opportunities for younger workers, who used to learn the trade in exactly those roles.
What to do if you hire in an exposed field
- Reprice the role with this year's number, not last year's. If your offer for a data, marketing or development role is still anchored to pay from two or three years ago, you lose candidates to offers that are now 46% higher.
- Separate the real raise from the composition effect. Before pricing a role as "senior because AI demands it," check whether the job genuinely needs that experience, or whether you are just following a posting trend and overpaying for a profile you do not need.
- Fund your own entry-level pipeline. If the market stops advertising junior roles in your field, whoever keeps training them, through internal programs, internships or apprenticeships, will have the senior staff in five years that everyone else failed to develop.
- Weigh the raise against your sector's layoff exposure. A strong pay increase is not a guarantee of job security; if your field also ranks high in AI-linked layoff announcements, say so plainly when hiring instead of promising a stability the data does not back.
On our own team, we keep training junior people alongside those who already have years in the trade, precisely because a team that only hires seniors stops having anyone to promote into that role a decade from now.