Are recruiters and job seekers really adopting AI in hiring?

LinkedIn's 2026 talent release makes one claim with two sides: AI is moving into hiring from the recruiter's side and from the candidate's side at once. On the recruiter side, "93% of recruiters say they plan to increase their use of AI in 2026, and 59% say it's already helping them discover candidates with skills they wouldn't have found before." Two-thirds of recruiters (66%) plan to increase AI use for pre-screening interviews, which "70% believe will help them have more valuable conversations with candidates." On the candidate side, "81% of people have or say they plan to use AI in their job search and nearly half (48%) say AI tools boost their interview confidence."

The excerpt frames this as a scarcity problem. "LinkedIn data shows US applicants per open role have doubled since the spring of 2022," and two-thirds of recruiters (66%) say qualified talent has become harder to find, under pressure "to fill roles faster (42%)" and to "uncover 'hidden gem' candidates (39%)." The AI products are presented as the answer to that squeeze at both ends. LinkedIn's job search lets people "search in your own words" without knowing keywords or titles, and its recruiter agent, Hiring Assistant, is said to help recruiters "uncover hidden gem candidates and save hours per role." The reasoning is the platform's own account of why the rollout makes sense, not an outside explanation of why adoption is happening.

The closest neighbors measure different things. The vacancy-data note (Is AI creating common skills across jobs or deepening divisions?) reads what employers ask for in job postings across ten countries. This excerpt reports what hirers and candidates say they will do with AI, which is adoption, not labor demand. The WORKBank note (What collaboration level do workers actually want with AI?) measures what workers want AI to do in their jobs. This release measures use, and says nothing about whether candidates want to be screened by an AI. The sharpest contrast is with the agent benchmark (Why do AI agents fail at workplace social interaction?), which tests agents on simulated workplace tasks and finds social interaction among the hardest. LinkedIn's AI-powered interviews and Hiring Pro, which helps hirers "reach out to the most relevant talent," are agents operating in that same kind of interaction, though the benchmark's failures were with colleagues, not candidates. The release reports no completion or failure rates for them.

What the excerpt does not establish is how the figures were gathered. It names no survey firm, sample, question wording or fieldwork method, and it offers no independent check on them. "Say" marks these as self-reports. The 81% mixes people who already use AI with people who only plan to, and the 59% and 70% are beliefs about helpfulness, not measured outcomes. The defensible claim is narrow: LinkedIn's own figures show recruiters and job seekers both expanding AI use in hiring. Whether that use improves matches is outside what this excerpt can show, and the sibling note takes up that gap.