If you think AI hiring is only for coders, Indeed’s new data is going to rearrange your mental model fast. The share of Indeed job postings containing AI terms has “skyrocketed over the past year,” and a focused look at AI-touched job titles shows a dramatic jump in the US from 264 titles in 2022 (2.6% of titles with at least five job postings) to 822 titles in the first quarter of 2026 (8.3%).

What makes this more than a headline number is the composition. Sixty-three percent of AI-touched US job titles were non-tech in the first quarter of 2026. In other words, employers are putting “AI” in job titles for roles that have existed for decades, not just building new AI-specialist functions.

To understand what “AI-touched” actually means in this analysis, the methodology matters. Pawel Adrjan, senior director of economic research for Europe, the Middle East, and Africa at the Indeed Hiring Lab, examined job titles defined as those with at least five postings mentioning AI in the title in a specific quarter. That definition keeps the focus on demand patterns that are visible in hiring volume, not one-off experiments.

The pattern Adrjan describes is the key operational implication for anyone hiring, training, or planning workforce strategy. Software development still had the highest share of AI-touched job titles in the first quarter of 2026, just like in 2022. But that share declined as other occupation types increased their share of AI-touched titles. Management, marketing, education, and instruction were among the occupational categories that increased their share.

That shift suggests a specific employer preference: field expertise plus AI fluency. A physical therapist using AI for documentation would still need to be skilled at helping patients. The job does not become “generic AI work.” Instead, AI becomes an input to the core work, and the title starts reflecting that combined requirement.

Adrjan makes the broader comparison in terms anyone can grasp: this is similar to the earlier broadening of computer skills beyond traditional IT roles. “One pattern that stands out is that many of the roles with AI in the title are jobs that have existed for decades,” Adrjan’s report said. The point is not that employers suddenly want everyone to become an AI specialist. It is that employers are adding AI to titles where using AI tools is required, which is “an indication of how AI is already reshaping jobs.”

The examples in the analysis make the trend feel real, not theoretical. They include “AI Autonomous Truck Test Driver,” “Physical Therapist (AI Documentation),” “AI Project Engineer,” and “Electrical Engineer - Battery Systems for AI Data Centers.” Those titles are basically job descriptions in disguise. They imply the employer expects workers to understand AI systems enough to operate them or document them, while still delivering the domain outcome the role has always demanded.

This matters because the common fear is replacement. Adrjan’s interpretation of the data points the other way. When a job title includes AI, the signal in the data looks more like demand than replacement. “It really seems to capture employers who are wanting AI skills to be incorporated into the job,” Adrjan said, “which looks a bit like augmentation.” Mentions of AI in job titles do not necessarily mean job seekers need a computer science degree or deep technical knowledge. Instead, Adrjan says employers want people with expertise in their field and AI fluency.

That framing is “reassuring” for people who worry about AI because the expectation is about applying AI to work they already know, in domains they already know, rather than switching into something completely different. For employers and boards, this is a subtle but important nuance: workforce plans may not need to fully reboot headcount, but they do need to update role definitions, training pathways, and hiring filters.

The analysis also intersects with a second Indeed Hiring Lab report by Guillermo Gallacher, an economist at Indeed Hiring Lab, who examined US job postings based on how exposed occupations were to AI. Gallacher said that “the more exposed to AI an occupation is, the more it declined” between May 2022 and May 2026. But in the shorter span from May 2025 to May 2026, Gallacher said “the more exposed to AI an occupation is, on average, the more it rebounded.”

Gallacher’s framing flips the relationship from job destruction to job creation, at least across those windows. Adrjan’s read of that pattern is that AI augmentation is more likely than AI creating entirely new categories of work by default. That still leaves an execution problem: if AI competence becomes an expectation across more occupations and more jobs, some workers may not be able to get training or familiarity with tools as fast as others.

Adrjan notes that people can upskill on their own, but employers and educational institutions can also incorporate AI training, especially when jobs require specialized AI tools. For decision-makers, the second-order risk is inequity of readiness: uneven AI tool access can translate into hiring and promotion friction, not because the roles disappear, but because the qualification bar quietly rises. The practical takeaway for peers is simple but urgent: AI is being baked into job titles across non-tech functions, and the organizations that treat “AI fluency” as a workforce requirement rather than a software-only topic will be better positioned as demand keeps spreading.