If you’ve recently filed an injury claim and gotten a settlement offer that felt insultingly low, you might be wondering whether a human being ever actually looked at your case. Increasingly, the answer is no. An algorithm did. And as of March 2026, state regulators are finally starting to ask hard questions about what that means for you.

The National Association of Insurance Commissioners launched a 12-state regulatory pilot last March, the first examination of its kind, specifically targeting how insurers use artificial intelligence to make injury claims decisions. A vote on nationwide rollout is scheduled for the NAIC’s November 2026 fall meeting. That timeline matters because the system being scrutinized is already making decisions that affect real people right now, today, while the oversight framework is still being built around it.

The Scale of What’s Already Happening

Here’s what I tell people when they’re surprised that a computer handled their claim: this isn’t a fringe practice anymore. By 2026, an estimated 91% of insurance companies have adopted AI technologies in some form. The AI-in-insurance market is projected to reach $35.8 billion by 2029. This isn’t experimental. It’s the industry’s new operating model.

The technical term you’ll start hearing is “straight-through processing,” which just means a claim gets resolved automatically from start to finish, no human adjuster involved. For simple claims, those rates have jumped from roughly 10 to 15% a few years ago to as high as 70 to 90% today. That’s a staggering shift. What it means practically is that the majority of straightforward claims are now being evaluated, valued, and settled by software.

For a minor fender-bender with a clean repair bill, that speed can genuinely be a good thing. You get paid faster. There’s less friction. But the moment your claim involves anything harder to quantify, a soft-tissue injury, emotional distress, ongoing treatment, future medical needs, the algorithm’s efficiency becomes your problem.

Why Soft-Tissue and Psychological Injuries Are Especially Vulnerable

Recommended AMORNING 276-Piece First Aid Kit

Portable first aid kit with 276 pieces covering wounds, sprains, and basic trauma care.

Check Price →
Injury TypeAI Valuation PatternDocumentation Challenge
Minor property damage (fender-bender)Faster processing, accurateRepair estimate quantifiable
Soft-tissue injury (whiplash, back pain)Systematically lowerDoesn’t appear on imaging
Psychological injury (anxiety, PTSD)Systematically lowerNo objective measurement
Ongoing/future medical needsUnderweightedRequires projection, not historical data
Pain affecting daily lifeUnderweightedSubjective, not scan-visible

Helpful resource: How to Win Your Personal Injury Claim by Joseph Matthews (Nolo) is a top-rated option for this. (As an Amazon Associate this site earns from qualifying purchases.)

This is the part nobody tells you upfront. Consumer advocates have documented that AI settlement platforms used by major insurers are specifically calibrated in ways that produce lower valuations for injuries that don’t show up cleanly on imaging or a repair estimate. Whiplash. Chronic back pain. Anxiety after a serious crash. These are real harms, but they’re hard to reduce to a number a model can process with confidence, so the model discounts them.

Think about what an algorithm is actually doing. It’s trained on historical claims data, which means it’s learning what past settlements looked like. If insurers historically lowballed soft-tissue claims (and they did, I know because I spent 12 years on that side of the table), the model learns to replicate that pattern and optimizes it further. It’s not malicious. It’s math. But the output is the same: an offer that doesn’t reflect your actual situation.

The injury picture makes this worse. Industry data from 2026 shows bodily injury severity in auto claims rose 9.2% year-over-year. People are getting hurt more seriously, or at least the documented harm is increasing. At the exact same time, AI tools are being deployed to speed up and cap settlement offers. That’s a widening gap between what injured people actually need and what the system is designed to give them.

What Regulators Are Actually Examining

You might be wondering what the NAIC pilot is specifically looking at. Based on reporting from Auto Body News in March 2026, the examination targets how AI systems influence total-loss determinations and claims payouts. That’s significant because it’s not just auditing whether AI is being used, it’s scrutinizing the outputs and the calibration behind them.

The NAIC has also taken a clear public position that existing insurance laws apply equally to AI-driven decisions. If a human adjuster made a bad-faith decision, the insurer is liable. If an AI system makes the same bad-faith decision, the insurer is still fully liable, including when that AI comes from a third-party vendor. That’s an important legal point. The insurer can’t point at software and shrug.

What’s still being worked out is the enforcement mechanism. Knowing the law applies is different from having examiners who understand enough about model architecture to spot when a system is systematically undervaluing a specific injury category. That’s the technical gap the pilot is trying to address before the November vote.

What This Means When You’re Sitting Across From an Offer

Here’s what I tell people in practical terms. The first offer from an AI-assisted system is a starting point, not a verdict. You have the right to understand how your claim was evaluated. Under most state insurance regulations, you can request documentation of how a settlement figure was determined. Whether the insurer’s response will actually explain the model’s logic is another question, and one regulators are now pressing.

If your injury involves ongoing treatment, future care, or pain that doesn’t show up on a scan, those elements almost certainly weren’t weighted appropriately by an automated system. That’s not speculation. It’s a documented pattern that reporting from Insurance News Net and consumer-facing legal sources like the California Accident Attorneys Blog have both described in detail in 2026.

Getting an independent medical evaluation, keeping meticulous records of how your injury affects daily life, and documenting every symptom, appointment, and cost creates the paper record an algorithm can’t produce on its own but that a human reviewer or attorney can use to push back. A personal injury attorney who understands how AI valuation tools work, and that’s becoming a specialized competency, can challenge low algorithmic offers in ways that weren’t necessary five years ago.

I can’t tell you what your specific claim is worth or whether you have grounds to dispute an offer. That’s a conversation for an attorney who knows your state’s laws and your particular facts. What I can tell you is that a low number from an AI system is not final, and the regulatory scrutiny now building around this practice suggests you’re not alone in thinking something feels off.

The November 2026 NAIC vote won’t fix the system overnight. But it will likely create new examination standards that, over time, give injured people better tools to challenge decisions made by software that was never designed with their interests in mind. In the meantime, the best protection is knowing the system exists and understanding that you have options beyond accepting the first number a computer generates about what your pain is worth.

Sources

This article is for general informational purposes only and does not constitute legal advice. Laws vary by state. Consult a licensed personal injury attorney in your jurisdiction for advice specific to your situation. Most personal injury attorneys offer free consultations.

Disclosure: As an Amazon Associate, we earn a small commission from qualifying purchases at no extra cost to you. We only recommend products that genuinely support the topics covered in this article.