--- title: Americans Would Let AI Run Their Job Search, but Few Know It Can Be Biased description: 81% of Americans would let AI handle part of their job search, but most have never heard the advice can be biased, a new Harris Poll survey finds. author: Darie Nani (Editor-in-Chief) updated: 2026-07-25T11:31:50.220Z canonical: https://www.sovereignmagazine.com/article/ruth-ai-harris-poll-job-search-survey-trust-gap image: https://cdn.nanimediahouse.com/ai-job-search-trust-gap-30981.webp categories: Artificial Intelligence content_type: News region: United States publication: Sovereign Magazine schema_type: Article --- Four in five American adults, 81 percent, would let an AI agent handle at least one part of their job search, according to a national survey of 2,131 US adults carried out by The Harris Poll in collaboration with Ruth AI, an AI career platform for women, in June 2026. Nearly half, 47 percent, would go further and let AI negotiate their salary. The numbers describe a fast shift from asking AI for career tips to handing it real decisions. Two-thirds of respondents said they would let AI search for open roles, and the same share would let it run pre-interview research, at 67 percent each. Sixty-five percent would let it update a resume, and 55 percent would let AI apply for jobs outright. Forty-nine percent would let it negotiate benefits such as vacation days or remote-work terms. Delegation is heaviest among younger workers. Ninety percent of Millennials would hand AI at least one job-search task, and 45 percent of all adults have already used a general-purpose model such as ChatGPT, Claude or Gemini for career or work advice, a figure that rises to 66 percent of Millennials and 63 percent of Gen Z. One in three adults has asked AI for advice specifically about money at work: salary, raises, bonuses or how to negotiate pay. ## A Trust That Runs Ahead of the Warnings Three in four adults, 76 percent, said they had never heard that independent research has found AI can produce biased career and salary guidance. Separately, 72 percent agreed AI can sound confident even when its advice turns out to be wrong. The gap is wide: many of the people willing to let a chatbot negotiate their pay are unaware that a chatbot's advice on that exact question has been shown to shift depending on who is asking. ## What the Research on AI Career Advice Shows The clearest evidence sits in a 2025 paper by Sorokovikova, Chizhov, Eremenko and Yamshchikov, which tested how large language models respond to identical prompts under different assigned personas. [Evaluating models with pre-prompted personas on a general-knowledge benchmark produced only small, mostly random differences in scores](https://arxiv.org/abs/2506.10491). Asking a model to grade a user's own answer showed more significant signs of bias. Salary negotiation advice showed the most pronounced bias of the three: the researchers found the guidance shifted by the asker's inferred identity once a model was asked to weigh in on pay. AI tools built for the workplace carry other documented bias problems of their own. [Automated screening systems have been found to disadvantage neurodivergent job seekers](https://www.sovereignmagazine.com/article/the-hidden-bias-in-ai-hiring-how-automated-screenings-are-failing-neurodivergent-job-seekers), and hiring algorithms trained on historical pay data can quietly reproduce the gaps they were meant to remove. Ruth AI, founded by Valerie Chapman, was built to work the other side of that problem, aiming to help women negotiate pay and plan careers with AI support, on a stated mission to close the gender wage gap. ## Who Is Leaning on AI Most Gen Z's use of AI for career decisions builds on habits already visible elsewhere. [The same generation has already turned to platforms like WorkTok for quick professional guidance](https://www.sovereignmagazine.com/article/worktok-gen-z-career-discovery), and the survey shows it is now just as willing to hand salary and job-search conversations to a chatbot. > "We are delegating faster than we are asking questions. The responsibility now falls on the people building AI to earn the trust users are already giving it." > — Valerie Chapman, Founder and CEO, Ruth AI The survey was fielded online from June 11 to 13, 2026, among a nationally representative sample of 2,131 US adults, including 420 Gen Z, 620 Millennial, 519 Gen X and 572 Baby Boomer respondents, weighted to the general adult population. Find out more at [ruthapp.ai](https://ruthapp.ai). ## FAQ **Q: Is it possible for AI to be biased in career advice?** Yes. A 2025 academic study found large language models gave measurably different salary negotiation advice depending on the persona of the person asking, even though general-knowledge answers from the same models showed little bias by comparison. **Q: Should I trust AI for career advice?** AI can be useful for research and drafting, but the survey above found most people are unaware its salary and negotiation guidance can carry bias. A specific pay recommendation from a chatbot is worth checking against outside sources rather than treated as neutral fact. **Q: Can I use an AI agent to apply to jobs?** Yes. Tools now exist that search listings, tailor a resume and submit applications automatically, and 55 percent of Americans surveyed said they would let AI do this. The same survey found awareness of its limits lags behind adoption.