Job Seekers Have Lost Their Leverage, and the Data Proves It
Sign Up For Daily Job Market News
Job seeker confidence just posted its worst quarter of the year, for a simple reason: more than half the people actively searching for work right now don’t have a job to fall back on.
ZipRecruiter’s Job Seeker Confidence Index fell 6.3 points in the third quarter of 2026, landing at 93.5. The survey reached more than 1,500 job seekers nationally between July 25 and August 12, and every part of the index moved in the same direction. The present situation reading dropped 6.8 points, preparedness fell 2.9, financial wellbeing dropped 7.6, and expectations took the hardest hit at 10.7 points, according to ZipRecruiter Research.
The number I keep coming back to is 51%. It’s the share of current job seekers who are unemployed, up from 41% in the first quarter. Only 38% of the previously employed group quit by choice, down from 43%, and 23% are searching because of a layoff. Unemployed candidates behave differently than employed ones, and most of what this survey found traces back to that single shift.
Pressure to take the first offer climbed to 67%, up from 63%, while willingness to negotiate slipped to 69%, down from 73%. Rent comes due before the second interview does, and the numbers reflect it. 40% of respondents describe themselves as financially struggling, and 55% believe employers hold the upper hand right now.
The gap that should bother both sides of the table: job seekers expect a median pay increase of 27% when they change jobs. Actual wage growth for job switchers is running 4.4% year over year. Someone in nearly every one of those conversations is going to be disappointed, and the candidate usually finds out last.
Pay still decides almost everything else in a search. 78% say it’s the top reason they accept an offer, 68% say it’s the top reason they reject one, and among people who already have a job, 73% say pay is why they’re looking elsewhere. If your comp bands haven’t been checked against real numbers recently, this is the quarter to do it.
For employers, the temptation in a market like this is obvious. Candidates have less leverage, so offers can come in lighter and negotiation gets easier to hold your ground on. I’d be careful leaning too hard into that. People remember what it felt like to get hired into a lowball offer, and anyone underpaid in a soft market starts looking again the day it turns.
The AI split in this data is quieter, and I think it matters more over the next year than the headline confidence number. 67% of job seekers already use AI in their daily work, and 43% list AI skills on applications, but only 10% make them prominent. Candidates are building the skill into their actual work long before it reaches the resume, so employers screening applications for AI experience are passing on people who already have it. This gap tracks with what we’re already seeing in the broader AI hiring data, where candidates using AI well are outperforming the ones who aren’t advertising it.
“Job seekers have navigated a lot of change over the past year and a half,” said Nicole Bachaud, labor economist at ZipRecruiter. “Job growth may have cooled off in Q2 leading into Q3, but they are adapting quickly. The next step is to turn their newfound AI-enabled skillsets into marketable skills on their resumes to capture the attention of employers.”
Looking six months out, only 25% of job seekers expect more jobs to open up, down from 30% in the first quarter, and 33% expect fewer opportunities, up from 27%.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
Jobless Claims Are Barely Moving, But the Average Just Turned
The Department of Labor released its weekly unemployment insurance numbers this morning, and on the surface, nothing changed. Initial claims came in at 203,000 for the week ending August 22, down 4,000 from a revised 207,000 the week before. Continuing claims fell to 1,778,000, and the insured unemployment rate held at 1.2%.
Both readings sit below where they were a year ago. Initial claims ran 229,000 in the comparable week of 2025, and insured unemployment was 1,942,000 with a 1.3% rate, according to the Department of Labor’s weekly claims report.
The number I’d actually watch is the 4-week moving average, which rose to 205,500, up 1,250, even while the weekly print fell. Weekly claims bounce around for reasons that have nothing to do with the economy. The average is what tells you when something real is changing, and right now it’s drifting up slightly rather than down.
State by state, there’s no sign of a layoff wave forming. Not a single state posted an initial claims increase above 1,000 this week. Michigan posted the largest decrease at 2,446 fewer claims, credited to fewer manufacturing layoffs, followed by California, South Carolina, and Pennsylvania. Federal civilian employee claims fell to 390, and newly discharged veterans filed 385, both down from the prior week.
Set this next to the ZipRecruiter numbers above, and you get a market that’s harder to describe in one sentence than usual. Claims measure separations. Job seekers experience hiring. Those are different questions, and both answers can be true at once.
Employers have mostly stopped letting people go, and at the same time, they’ve slowed down bringing new people in. For anyone already holding a job, this is one of the safer stretches in years. For anyone trying to get in the door, it’s about as difficult as it’s been outside of 2009 and 2020.
If you’re one of the employers still actively hiring in this environment, that gap is your advantage. Candidates are motivated, available, and easier to close than they were eighteen months ago.
Workers Are Turning on AI, and Their Bosses Haven’t Noticed
Glassdoor pulled every U.S. company review that mentions AI, LLMs, GPTs, artificial intelligence, or OpenAI, then sorted the mentions by whether they landed under pros or cons. The findings should worry any employer who thinks their AI rollout has been well received.
Mentions of AI in company reviews rose 240% from May 2025 to May 2026, on top of increases of 202% in 2023, 74% in 2024, and 164% in 2025, according to Glassdoor’s research, led by senior economist Chris Martin. The tone flipped hard along the way. In 2019, 81% of reviews mentioning AI framed it as a pro. In 2026, that’s down to 43%, and 53% of AI mentions now land only as a complaint.
Who feels good about AI at work depends almost entirely on the job. Company leaders are the most positive group, at 67%, followed by sales and product managers, recruiters, software architects, and data analysts. Insurance claims adjusters sit at the opposite end, with 98% of their AI comments negative. Writers, journalists, accountants, customer service reps, designers, and IT staff also skew heavily critical.
“Insurance claims adjusters are shockingly negative about AI, with 98% of comments being negative,” Martin said.
The complaints break into clear categories. 20% of negative mentions are about jobs being replaced, 14% are about employees being force-fed AI tools with no input on the rollout, 13% say AI is distracting from the core business, and 10% raise workplace surveillance concerns. The last three all trace back to how the tool arrived and who was consulted before it did, which is a rollout problem an employer can actually fix.
The number I’d put in front of a leadership team is 47%. Workers who describe AI as a pro in their reviews are 47% less likely to apply to other jobs on Glassdoor. On the flip side, 25% of workers who list AI as a con also mention a layoff in the same review, and AI-critical reviewers are 3.7 times as likely to mention burnout and six times as likely to mention layoffs. “A whopping 25% of workers who mention AI in the ‘cons’ section of their reviews also mention layoff,” Martin noted.
Confidence in leadership tracks the same line almost exactly. Among AI-critical reviewers, only 23% approve of the CEO and 18% report a positive business outlook. Among AI-positive reviewers, those numbers jump to 86% and 81%. There’s a gender gap too: men’s AI comments run 45% positive against 32% for women, and among Gen Z specifically, it’s 42% for men versus 21% for women.
For employers, the fix here has less to do with the technology and more to do with the rollout: treat it as a staffing and communication decision, with input from the people using the tools every day. Candidates read company reviews before they apply, whether or not anyone planned for that. A bad AI rollout has already become a recruiting problem, and it lands in the same review that tells the next applicant whether to trust your company.
