Insurance Claims Adjusters Are Pushing Back Against AI as Errors, Workload and Job Cuts Mount

Insurance claims adjusters are emerging as one of the most openly anti-AI professions in the U.S. workplace, according to fresh Glassdoor research that found 98% of posts mentioning AI from claims adjusters were negative. The backlash matters because insurers are increasingly using AI to triage claims, summarize records and even generate payouts, raising fears that the technology is creating more mistakes than efficiencies.

The findings arrive as claims employment is shrinking, entry-level openings are drying up and insurers continue to market automation as a way to speed up customer service. But adjusters say the reality on the ground is often more paperwork, more rerouting and more cleanup after systems get key details wrong.

For the industry, the dispute is about more than employee attitude. It is becoming a test case for whether AI can be introduced into a highly sensitive, high-stakes workflow without eroding trust among workers and policyholders alike.

Why are claims adjusters so critical of AI?

Claims adjusters are critical of AI because they say it is being pushed into their jobs faster than it can reliably handle the complexity of insurance work. When the technology misreads a file, misclassifies a claim or produces a flawed summary, the resulting problems usually fall back on humans to fix.

That frustration is visible in Glassdoor posts, where reviewers have complained that leaders are forcing AI tools on staff and clients even when the output is unreliable. In a profession built around accuracy, empathy and documentation, many adjusters say the technology often feels less like support and more like an extra layer of risk.

“There’s an AI fatigue,” said Geoffrey Conrad, a claims executive in Mobile, Alabama, describing what he sees as exhaustion from constant pressure to adopt new tools.

Chris Martin, a senior economist at Glassdoor, said he did not initially expect claims adjusters to register such strong resistance. But after seeing the data, he concluded the field is in the middle of a major transition, with workers reacting both to job insecurity and to poor technology rollouts.

What the Glassdoor data shows

Glassdoor’s analysis suggests claims adjusters are unusually skeptical of AI compared with other occupations. The company said 98% of AI-related reviews from claims adjusters were negative, making them the most hostile workforce segment in its dataset.

The criticism generally centered on AI being mandated from the top down. Reviewers said tools were often inaccurate, and that workers were expected to absorb the fallout when the systems failed. Martin said skepticism tends to spike when employees believe layoffs are ahead or when the product being forced on them is clearly underperforming.

The findings fit a broader pattern in workplace sentiment: people are more likely to reject AI when they see it as a substitute for judgment rather than a support tool. In claims handling, where a small mistake can affect payments, medical reimbursements or damage estimates, that line matters.

How AI is being used in insurance claims

AI is being deployed across insurance in ways that can directly shape how quickly policyholders are paid and how much attention a claim receives. Insurers are using chatbots to collect first notice of loss, computer vision to inspect damage from photos and video, and language models to summarize long medical and legal records.

In a typical AI-heavy workflow, a policyholder might upload images, documents and receipts. Software then classifies the incident, extracts relevant details and may even recommend a settlement or route the file to a human employee only if the system flags uncertainty.

In theory, that approach speeds up simple claims while freeing humans to focus on difficult cases. In practice, adjusters say the technology often produces errors that require more manual work, not less.

• Initial claim intake through chatbots or automated forms
• Damage assessment from uploaded photos and video
• Document summarization for medical and legal records
• Claims routing based on automated classification
• Potential fast-track payouts for low-complexity cases

How did one adjuster’s work change after AI was introduced?

One former claims employee, Ahmad Jackson, said the technology was supposed to improve efficiency but instead created a flood of misfiled claims and inaccurate summaries. He worked in the claims department of a major insurer and saw the company introduce AI for initial loss reporting, the stage when a customer first reports an incident and the claim is opened.

Jackson said the system often placed claims in the wrong workflow, forcing staff to redirect them manually. He also encountered AI-generated hallucinations in claim summaries, which created a new kind of risk: if he repeated the faulty information to a claimant or an attorney, the error became his problem.

He left that job and moved to another insurance carrier. In his view, the technology was not reducing pressure on adjusters; it was shifting the burden back onto them.

Jackson said the tools were “getting things wrong” and were “implementing more work onto the adjusters.”

Why insurers still push automation

Insurers and insurance startups are still investing heavily in AI because the economics are compelling if the systems work as advertised. Claims handling is labor-intensive, repetitive in many cases and often delayed by paperwork, making it an attractive target for automation.

Startups such as Liberate and Pace have raised capital by promising to reinvent the insurance process with AI. Established players, meanwhile, are also widening their use of automated tools to improve speed, cut overhead and handle basic claims without requiring a person for every step.

Lemonade has become the clearest example of that strategy. Since its launch in 2015, the company has pitched itself as a digital-first insurer built around bots and machine learning rather than traditional bureaucracy. By the end of last year, its AI chatbot, AI Jim, was handling initial reports 96% of the time, and automation was handling about 55% of all claims.

A company spokesperson, Paul Staats, said automation allows staff to concentrate on the most difficult matters and bring their empathy and expertise to complex claims. He also said the Glassdoor findings should be taken seriously by the wider insurance sector.

According to Staats, AI can help workers reserve “empathy, care, and expertise” for the hardest cases.

How severe is the job decline in claims handling?

The employment picture helps explain why AI adoption is so controversial. The U.S. Bureau of Labor Statistics projected in 2024 that claims adjuster jobs would fall by 18,900 over the following decade, a decline of about 5%. More recent data suggests the contraction may already be happening faster than expected.

Between May 2025 and May 2026, BLS data showed employment in the sector dropping 21%. Glassdoor also reported that entry-level postings in the field were down 50% since 2025, suggesting that younger workers may be seeing the sharpest impact first.

For workers, those numbers feed a sense that automation is not merely changing tasks but shrinking the profession itself. Martin said the lack of visible career prospects can drive even more negative reactions to AI, especially when employees think the tools are there to justify staffing cuts.

What do adjusters say AI gets wrong?

Adjusters say AI errors are especially damaging because even small omissions can change the outcome of a claim. A smudge on a document, a missing note in a medical summary or an incorrectly parsed attorney letter can all lead to a bad payout decision, a delayed claim or a customer dispute.

Sandy Avina, a former claims adjuster who now works as an insurance industry consultant, said adjusters generally lack confidence in AI output because the systems can miss context that a human would catch immediately. That includes details buried in medical files or legal correspondence that are easy for a model to overlook or misread.

The customer often does not see the source of the mistake. Instead, the policyholder usually assumes the adjuster made the error, even when the fault began with an automated summary or classification step.

When AI helps, and when it doesn’t

Some adjusters do use AI for limited, low-stakes administrative work. Jackson said he found it useful for routine tasks such as extending a rental car reservation or managing other nuisance calls that do not require much interpretation.

That distinction is important. Many workers are not opposed to every AI use; they are opposed to handing over judgment-heavy decisions to systems they do not trust. In other words, the resistance is not to assistance, but to replacement.

Conrad said AI should be treated as a tool, not as a decision-maker, arguing that it should never be “given the keys.”

What makes claims work different from other AI jobs?

Claims work is different because it combines customer service, document review, financial judgment and emotional support in a single role. A claims adjuster may be helping someone after a fire, a car crash, a flood or a medical emergency, which means accuracy is only part of the job; tone and trust matter too.

That human side is one reason automation has drawn such a strong reaction. In catastrophe situations, policyholders often want reassurance as much as a payout estimate. A system that can process images quickly may still fail to provide the sense that a person is listening.

Conrad spoke from personal experience. Years before his insurance career, he lost his home in a fire and described the event as a total loss. For him, the idea that an algorithm might replace a person’s on-the-ground judgment during such a moment is unsettling.

Conrad said that if someone had told him then that an estimate would be produced from photographs alone, he would have been furious, because what he needed most was confirmation that he and his family were safe.

What the industry debate means now

The controversy over claims-adjustment AI is likely to spread beyond insurance because it captures a broader issue facing white-collar automation: the difference between efficiency and accountability. A tool that saves time in one part of the workflow can still create downstream costs if it fails quietly and frequently.

That is why the insurance sector is such a revealing case study. Claims handling is not a flashy frontier for AI, but it is exactly the kind of operational process where companies are tempted to automate first. The stakes, however, are high enough that every error can have financial and emotional consequences.

For insurers, the challenge is not just to make AI work, but to make employees and customers trust it. If workers believe the technology is being imposed to reduce headcount, they are likely to keep resisting it. If customers believe the system is making life harder after a loss, the brand risk rises fast.

For now, the evidence suggests many adjusters think the industry is moving too fast. They are willing to use AI where it saves time, but not where it replaces judgment, context or empathy.

Key timeline of the claims-AI shift

The insurance industry’s turn toward AI has accelerated over several years, but the backlash is becoming more visible as adoption deepens.

Bottom line for insurers and workers

Claims adjusters are not rejecting AI because it is new; they are rejecting it because, in their view, it is being deployed before it is ready. The strongest criticism is not that AI has no place in insurance, but that it is too often being used in situations where errors are expensive and human reassurance matters.

As insurers keep automating intake, summaries and low-level decisions, they will need to prove that AI can actually reduce friction rather than simply relocate it. Otherwise, the profession’s backlash may be less a temporary adjustment and more a warning sign for other industries moving fast on AI.