Careers

Asked back after an AI layoff: what to request on pay, security and transparency

A new survey says nearly one in five US respondents were replaced by AI and invited to return, but the evidence on what returners actually win is thin.

12 min read
Empty bright meeting room with a white table, white chairs, a blank wall-mounted whiteboard and a light wood floor.
Image by Pexels from Pixabay

A Fortune report on an SHL survey says 19.1% of US respondents were replaced by AI and later asked back. This article separates what that figure measures from what is known about pay, written job protections and AI disclosure for workers offered a return.

Nearly one in five American respondents to a 2026 survey say AI took their job and the employer later asked them to return. According to a Fortune report published Nov. 9, 2026, the survey by the talent-assessment firm SHL found that 19.1% of US respondents said they had personally been replaced by AI and then asked back. About half of those asked back said they wanted a substantial raise before agreeing.

The headline number is softer than it sounds, and evidence on what returning workers actually obtain is close to nonexistent. What exists is a mix of stated preferences, employer surveys and sample contract language. Taken together, these show where a returning worker's leverage lies and which terms are worth getting in writing: pay, job security and disclosure of how AI is used in the role.

What the 19.1% figure does and does not mean

As summarized by a DEV Community write-up, SHL commissioned Propeller Insights to survey more than 1,000 US workers and more than 1,000 UK workers between June and July 2026. The US result was 19.1%. The same summary reports 16% for the US and UK combined and 24% among respondents aged 18 to 24.

Several limits apply. The data are self-reported, and "replaced by AI" reflects the worker's own account or the employer's stated reason, not verified causation. SHL sells talent assessment, so it has a commercial interest in the topic. Sample design, margin of error and question wording were not available for review.

The denominator is also unresolved. The Fortune headline says "nearly 20% of U.S. workers replaced by AI," but the survey figure is described as 19.1% of American respondents who said they had been replaced and asked back. One secondary summary phrases it as a share of workers who say AI replaced their job. It is not safe to read the number as a measured recall rate among everyone laid off for AI.

The reported demands carry a similar caveat. They describe what returnees said they wanted before agreeing to come back, not what employers granted. The table below shows the figures as reported by the DEV Community summary of the SHL survey; Dataconomy independently reports the 51% figure, though both trace to SHL.

What workers asked back after AI replacement said they wanted, US, June–July 2026 (SHL survey, self-reported, more than 1,000 US respondents)
What returnees wanted before agreeing Share
Substantial pay increase 51%
Written job-security guarantees 42%
Employer says plainly where AI is used 40%

Why employers are calling people back

Employer surveys point the same direction

Several surveys of employers, run by different firms, suggest that regret and rehiring after AI-driven cuts are not rare. Careerminds, an outplacement firm, surveyed 600 HR professionals who had run layoffs in the prior 12 months in February 2026. According to a Forbes Council contributor post, two in three companies that made AI-driven cuts were already rehiring, and 52% said rehiring began within six months.

Programs.com, citing the same survey, reports how the economics turned out. Some 30.9% of organizations said rehiring cost more than the layoffs had saved, 42.4% said the two roughly canceled out, and 26.7% ended up ahead. That split is the clearest evidence for worker leverage in the brief: for roughly three in four organizations, the cuts did not pay off cleanly.

Other surveys land in a similar range, with different definitions. Orgvue reported in March 2026 that 32% of organizations that dismissed workers because of AI later had to rehire; separately, 39% of 1,000 senior decision-makers had made AI layoffs and 55% of those said it was the wrong choice. Robert Half reported in July that 32% of US hiring managers said the same or similar roles were reinstated after AI-driven cuts, while another outlet gives 29%, so neither figure is verified.

Forrester's 2026 predictions, as reported by The Next Web, say 55% of employers regret AI-related layoffs and that half of AI-attributed layoffs will be quietly reversed. Some outlets attach the 55% figure to Orgvue instead, so the attribution is unresolved. Gartner, according to a Forbes Council post, forecasts that half of companies that credited AI for headcount cuts will rehire for similar functions by 2027; that is a forecast, not a measurement. The same post cites Visier data on 2.4 million workers at 142 companies showing the highest rate of former employees being rehired since 2018.

Company examples are contested

Klarna is the most cited case. In February 2024 the company said its OpenAI-based assistant was doing the work of 700 agents. In May 2025, according to Forbes, its CEO told Bloomberg that cost had been too dominant a factor and that quality had suffered.

Klarna disputes the "reversal" label. It said the pilot involved two new agents in a flexible, remote setup, and that this was an addition rather than a rehire. A tracker citing the company's 2025 annual filing says AI handled 80% of chats, which points to hybrid staffing rather than a mass recall. Whether laid-off agents were the ones brought back is not shown.

A Fortune syndication reports that Ford hired and promoted 350 experienced engineers in June after AI-powered quality-control tools underperformed. Another outlet says "reinstated or promoted," so the exact wording is unclear. These cases show that rebuilding human capacity happens at named companies, but they do not show what the returning people were paid.

The backdrop: AI-cited cuts are large, but "cited" is not "caused"

Challenger, Gray & Christmas counted 121,401 AI-cited announced job cuts in the first ten months of 2026, about 20% of all announced cuts, according to a report on the October data. The 2025 figure was 54,836. Challenger cautions that these are announcements rather than completed layoffs, and that the reasons are what employers say publicly. A company citing AI does not prove the technology replaced those workers.

Per CoinCentral's account of the Nov. 6 report, weaker market and economic conditions overtook AI as the leading stated reason for layoffs. That matters for a returning worker: if AI was partly a label for cost cutting, an employer's offer to come back may reflect a business problem that predates the technology.

Pay: a wish for a raise versus reports of lower wages

The two available signals on pay point in opposite directions. In the SHL survey, 51% of workers asked back said they wanted a substantial pay increase before agreeing. Forrester, as reported by The Next Web and others, says reversed jobs often come back offshore or at significantly lower wages.

These do not contradict each other because they measure different things. SHL captures what workers wanted. Forrester describes employer behavior. Neither measures realized pay for people who actually returned, so no source in the research supports the claim that returners get raises.

Leverage on pay therefore rests on indirect evidence. Employers that rehire have already absorbed a costly reversal, and the Careerminds survey found that nearly 31% spent more than the cuts saved. An employer in that position has reasons to prefer a known, trained worker over a fresh search. That is an inference from employer-side data, not a finding about wages.

Practitioner sources and sample contracts suggest the most concrete pay asks are specific ones: the base rate, and an explicit comparison with the prior rate. Title and scope belong in the same conversation. A secondary source in the research notes that rehires often come back under new job titles, and Forrester warns that firms may use AI to absorb simpler tasks while refilling mid-career roles elsewhere. A returning role may therefore not be the role that was cut.

Job security: what can actually be put in writing

Of workers asked back in the SHL survey, 42% wanted written job-security guarantees. In practice, a promise of security is only as strong as its wording. Recall rights, where they exist, come from a union contract, a written employer policy or a specific state or local law. According to a secondary blog, they are not automatic in at-will private employment.

What real recall language looks like

Examples come from institutions rather than from universal rules. The University of California, Santa Cruz defines preferential rehire as first consideration for a vacant position at the same or lower salary range maximum. A sample collective agreement on Law Insider restores seniority for benefits purposes if a worker is rehired, unless the benefit plan's own terms limit it, and says declining a non-equivalent position can mean forfeiting recall rights for that position.

Severance clawbacks and tenure resets

A law-firm blog by Aaron Hall says many severance agreements include clawback or offset provisions. These can require repayment of severance, or let the employer offset future wages, if the worker is rehired within a set period. The same blog says rehires are often treated as new employees, so prior service may not count toward benefits or retirement milestones, and that a "bridge of service" agreement can preserve it.

The blog also says modern agreements may define re-employment broadly enough to include 1099 contracting. That source is a low-authority blog and not primary law, so these claims are best read as hypotheses to check against the actual severance agreement and, ideally, with an employment lawyer. This article offers researched context, not legal advice.

A checklist of terms drawn from the sources

The research supports a set of written terms that follow from the rehire-mechanics sources. They are drawn from sample contracts and practitioner commentary, not from outcome studies, and no study shows which of them succeed.

  • Base pay, with the prior rate stated for comparison
  • Title and scope of the role
  • Credit for prior service toward benefits and retirement milestones
  • How any severance clawback or offset will be treated
  • A defined notice or severance period if the role is cut again
  • A description of how AI is used in the role

AI transparency: mostly a contract ask, not a legal right

In the SHL survey, 40% of returnees wanted the employer to say plainly where AI is used. SHL's Lucy Beaumont said companies were making fast decisions about workers on an incomplete view of what employees could do with and without AI. That gap is the practical argument for the request: it is a question about how the work will now be divided between people and software.

US law covers only part of this. Illinois HB 3773 (Public Act 103-0804) took effect Jan. 1, 2026, according to law-firm and compliance summaries. It makes using AI with a discriminatory effect a civil-rights violation, and it also makes failing to notify employees of the employer's use of AI a violation. Coverage runs from recruitment through discipline and termination.

The details of notice are unsettled. Seyfarth reports that the Illinois Department of Human Rights published proposed rules on May 15, 2026 and withdrew them 18 days later. As of late September they had not been refiled. One site gives a different withdrawal date, so the date is disputed, but the statutory duty is in force while the form and timing of a compliant notice remain unclear.

Other jurisdictions add narrower duties. New York City's Local Law 144, enforced since July 5, 2023, requires notice at least 10 business days before AI is used for hiring and promotion decisions in the city. Colorado's SB 26-189 was signed May 14, 2026 and replaces the 2024 Colorado AI Act; a single compliance blog gives Jan. 1, 2027 as its effective date, which could not be confirmed elsewhere.

None of these laws, as summarized, gives a general right to know where AI is used in a team's work or on what basis a job was cut. For most returning workers, a transparency request is therefore a negotiating ask rather than a statutory entitlement. The research found no example of such a clause in an individual contract, so how employers respond is untested. These conclusions rest on legal summaries, not on the statute text.

Risks: different people, different terms

An offer to return can differ from the old job in ways the headline figures hide. Forrester warns that the people rehired may not be the laid-off domestic workers at all, because some reversals refill roles offshore or at lower pay. Careerminds and other employer surveys count rehiring, not who was rehired, so they cannot say how often former employees get the call.

Contractor status is a second risk. If a rehire is structured as 1099 work, the blog sources suggest that benefits, tenure and severance protections may not carry over, and a broad re-employment definition could interact with a clawback. This is flagged in the research as a point to verify, not an established pattern.

Younger workers may be the most exposed. The Stanford Digital Economy Lab, as relayed by Fortune, found employment among workers aged 22 to 25 in highly AI-exposed occupations was 19% below where it would have been had it grown like that of less-exposed peers. The research did not include the study itself. The SHL survey separately found a higher asked-back share, 24%, among respondents aged 18 to 24.

What to watch

The strongest claim the evidence supports is modest. Several independent employer surveys agree that a meaningful share of companies that cut for AI are rehiring or regret the cuts, with estimates running from about a third to more than half depending on definitions. That gives a returning worker a real bargaining position, though not a guaranteed raise.

Three developments will show whether that position translates into terms. One is the original SHL data, which could settle what the 19.1% really measures. Another is whether Illinois refiles its notice rules and whether Colorado's replacement law takes effect as one source says. The third is any evidence on realized outcomes, since until someone tracks what returners get, the most reliable advice is to get every promise in writing and read the severance agreement first.

Frequently asked questions

Are 20% of workers replaced by AI being asked back?

Not as a measured rate. The SHL survey reports that 19.1% of American respondents said they had personally been replaced by AI and then asked to return. The denominator is unconfirmed, the data are self-reported, and the survey was commissioned by a company that sells talent assessment.

Do workers who return after an AI layoff get raises?

The research does not show that. In the SHL survey, 51% of returnees said they wanted a substantial raise, but that is a stated preference. Forrester, as reported by The Next Web, says reversed jobs often come back offshore or at significantly lower wages.

Does a laid-off worker have a legal right to be recalled?

Not automatically in at-will private employment, according to a secondary blog. Recall rights generally come from a union contract, a written employer policy or a specific state or local law. Sample language from the University of California, Santa Cruz and a collective agreement shows what such terms can look like.

Can severance be clawed back if an employer rehires a worker?

A law-firm blog says many severance agreements include clawback or offset provisions triggered by rehire within a set period. The terms vary by agreement, so the worker's own document governs. An employment lawyer can interpret it.

Does the law require employers to disclose where they use AI?

Only in narrow cases. Illinois HB 3773 requires notice of AI use in covered employment decisions, though its implementing rules were withdrawn, and New York City's Local Law 144 requires notice for hiring and promotion. As summarized in the research, none gives a general right to know where AI is used in a team's work.

Sources (20)

  1. Nearly 20% Of U.S. Workers Replaced By AI Are Asked Back , Dataconomy (2026-11-09)
  2. Almost one in five workers AI replaced were asked back , DEV Community (voxahq)
  3. Employers are quietly rehiring the people AI replaced, and paying them less to come back , The Next Web (2026-08-23)
  4. Council Post: Why Companies Regret Laying Off Workers For AI , Forbes Technology Council (2026-04-24)
  5. AI Layoff Reversals: Every Company That Rehired Workers Replaced by AI , Programs.com
  6. AI Layoffs Backfiring: Companies Are Rehiring , Ban the Bots
  7. Klarna Reverses AI Push, Says Customers Prefer Human Support , Forbes (2025-05-18)
  8. Klarna resumes hiring human agents after AI-first customer service produced 'lower quality' , BankingNewsAI
  9. ‘AI rebounding’: Nearly 20% of U.S. workers replaced by AI are being asked back (syndication) , MSNBCTV.news
  10. US Companies Rehire Workers Replaced by AI as Automation Falls Short , The 420.in
  11. AI Layoff Gamble Backfires as 90% of Executives Report No Productivity Gains , CoinCentral
  12. Can You Rehire an Employee After Severance? , Aaron Hall
  13. Understanding Preferential Rehire and Recall Rights , University of California, Santa Cruz Staff HR
  14. Recall from Lay Sample Clauses , Law Insider
  15. WARN Act and Recall Rights: Can Laid-Off Workers Be Called Back? , DEV Community (LayoffAtlas)
  16. Illinois Department of Human Rights Temporarily Withdraws Proposed Rules on Use of Artificial Intelligence in Employment , Seyfarth
  17. Illinois Passes New Law to Address AI in the Workplace , Morgan Lewis
  18. Illinois HB 3773: An AI Notice Duty With No Rules Behind It , Casrai
  19. AI Hiring Laws: Illinois HB 3773, NYC Local Law 144, Colorado , Meilynx (2026-10-11)
  20. Illinois AI Law in 2026: What HB 3773 Requires of Employers , Fruggr

Share this article

Back to list