When Everyone Can Sue: How AI Could Break the Insurance Industry
For decades, insurers have relied on a familiar playbook: issue a boilerplate denial, assume most people won’t hire a lawyer, and move on. That calculus rested on two frictions: legal knowledge and legal cost. AI crushes both. Drafting a complaint, formatting exhibits, citing rules of civil procedure, and e-filing can now be coached step by step by a bot. When those frictions drop, more people push back.
State court leaders are already discussing this shift in plain terms. The National Center for State Courts (NCSC) 2025 report “Trends in State Courts” highlights AI’s promise for self-represented litigants and court services, noting that courts are preparing for increased pro se participation supported by technology.[1]
The Pro Se Baseline Is Already Large
Even before modern chatbots, self-representation was a fixture of the legal system. In 2023, nearly half of new filings in the U.S. federal courts of appeals (46%) were pro se: 18,517 out of 39,987 total filings.[2] Between 2009 and 2012, non-prisoner pro se civil filings increased each year, with real property disputes a notable driver.[3]
Courts are now encountering self-represented litigants using AI, sometimes clumsily. In 2024, the Missouri Court of Appeals sanctioned a pro se litigant $10,000 for submitting 22 fake AI-generated citations out of 24 total, an early example of the messy learning curve.[4] The NCSC’s 2024 “Trends” commentary warns that as such tools improve, the volume of self-generated pleadings will rise even faster.
Where Litigation Costs Hide in Insurance Pricing
To see how this bites insurers, follow the line items. In property-casualty insurance, defense and cost containment expenses (DCC) are a subset of loss adjustment expenses (LAE), alongside adjusting and investigation costs. Based on industry aggregates, DCC typically accounts for approximately 1.5 to 2 percent of premium, with adjusting and other costs contributing another 6 to 8 percent depending on the line of business.[5] These ratios fluctuate widely by state and product; they should be understood as approximate industry averages, not fixed values.
At the macro level, the National Association of Insurance Commissioners’ 2024 Mid-Year Industry Report showed net losses and LAE totaling about 72 percent of net premiums earned for property-casualty carriers, illustrating how sensitive the combined ratio is to any expense creep.[6]
Zooming in, S&P Global’s homeowners insurance analysis found net losses and LAE surged to $101.29 billion in 2023, up 21 percent year over year, while premiums grew 10.8 percent. That’s a widening gap between claims cost and revenue, even before AI-related litigation volume enters the picture.[7]
Early Warning Signs in the Data
Direct, longitudinal proof that AI is raising litigation costs doesn’t yet exist, but adjacent indicators are flashing. The Swiss Re Institute’s “Social Inflation: Litigation Costs Drive Claims Inflation” report finds that litigation-related “social inflation” increased U.S. liability claims by 57 percent over the past decade, peaking at 7 percent annual growth in 2023.[8] While this trend predates mass legal AI adoption, it demonstrates the structural vulnerability: once more claims proceed to litigation, defense costs climb steeply.
Consulting firms are seeing the same pattern. Milliman notes that defense inefficiencies and leakage consume a significant share of carrier spend, and warns that increasing claim frequency or complexity, including AI-aided filings, could intensify those costs.[9]
Courts, meanwhile, are designing AI systems specifically for self-represented parties. The NCSC projects these tools will guide filings, explain deadlines, and generate forms, effectively lowering the barrier to suit for denied claimants.[1] That correlation isn’t yet proven causation, but the groundwork for volume growth is visible.
A Simple Scenario: What Happens When “Slink Away” Becomes “Sue”
We can’t predict perfectly, but we can bound it. Based on industry data, defense and cost containment currently average around 1.5 to 1.7 percent of premium, with adjusting and other expenses roughly 7 percent. That yields an LAE band of about 8.5 to 9.5 percent, varying by line and jurisdiction.[5] Litigated claims are dramatically costlier than non-litigated ones; defense costs can reach tens of thousands per claim once counsel is engaged.[9] AI-assisted self-representation likely increases the proportion of denied claims that proceed to litigation or pre-suit negotiation.
Suppose AI tools convert just 2 to 3 percent more denied small business, homeowners, and health insurance claims into lawsuits or formal demand packages. Even if half resolve early, those that progress through motions add substantial defense hours. The result could raise DCC by 30 to 50 basis points of premium over a few underwriting cycles.
If defense and cost containment currently average 1.6 percent of premium, an AI-driven surge could push that toward 2.0 to 2.5 percent in exposed lines, a 25 to 60 percent increase in DCC that directly inflates combined ratios. These are illustrative estimates based on modeling assumptions, not observed forecasts, but they reflect the sensitivity of insurer economics to litigation frequency.
Why the Old Denial Playbook Stops Working
The old denial math (issue a generic letter and assume claimants will give up) depended on the difficulty and expense of suing. AI changes that. It writes the complaint, formats the caption, explains service rules, and drafts an opposition motion. It gives frustrated policyholders the courage and competence to file. Even imperfect pro se complaints trigger duty-to-defend obligations and billable hours.
Insurers are deploying their own AI in response. Expect automated triage, early settlement models, and demand package analytics. Some carriers already report savings: McKinsey & Company notes Aviva saved more than £60 million on motor claims in 2024 using AI-driven fraud and repair analytics.[10] Yet if claimant-side adoption outpaces defense automation, overall litigation costs still rise.
What Smart Insurers Will Do Next
1. Insurers will ultimately rethink the denial. Weak denials invite lawsuits. Clearer, defensible decisions and data-driven settlement offers may prevent disputes from escalating.
2. Invest in early resolution tooling. If a multi-thousand dollar defense trajectory can be replaced by a modest settlement and goodwill payment, the math favors early resolution, though specific cost figures vary widely by case type and jurisdiction.[9]
3. Price for volatility. Carriers should model a 25 to 60 percent rise in defense costs over 2 to 3 years for high-exposure lines, recognizing these are scenario-based estimates that will differ across products and markets.
4. Watch the courts. If courts launch AI filing assistants and guided forms, that’s a leading indicator of local defense cost pressure.[1]
The Bottom Line
AI gives ordinary people enough procedure and structure to get into court. When more denied claimants do, defense dollars rise. Combined ratios are highly sensitive to small movements in LAE, and even minor increases in litigation frequency can ripple through pricing and availability.
If just a fraction of the “slink away” population becomes the “sue and pursue” population, insurers’ defense costs could jump from roughly 1.6 percent of premium into the mid-2s, a seemingly small change that would erase billions in underwriting margin. The old denial economy isn’t built for a world where everyone has a litigator in their pocket.
My Take
The insurance industry won’t be destroyed, but it will be forced to change in a world where everyone can litigate thanks to AI. That is a good thing. Anything that improves access to justice is a step forward.
A justice system that demands truckloads of money just to get a fair hearing is fundamentally broken. When people cannot afford to challenge unfair treatment, it invites abuse, especially in the form of shady insurance denials.
I am not anti-insurance. I am pro-insurance. I carry homeowners, health, life, auto, and an umbrella policy for extra coverage. But that does not mean I am pro-shady practices, and I have seen enough of them in my years practicing personal injury law to know how they work.
AI finally levels the justice playing field. I do not love the idea of premiums rising, but if that is the cost of fewer wrongful denials and more honest settlements, it is a trade-off worth making. The irony is that insurance companies will change, not because of ethics, but because AI will make it too expensive not to. That may be the truest form of accountability there is.
References
- National Center for State Courts. Trends in State Courts 2025. July 2025.
- Administrative Office of the U.S. Courts. Judicial Business 2023. Available at: uscourts.gov.
- Administrative Office of the U.S. Courts. Just the Facts: Trends in Pro Se Civil Litigation (2000-2019). Washington, D.C., 2021.
- Missouri Court of Appeals, Eastern District. Kruse v. Karlen, No. ED111172 (Mo. Ct. App. Feb. 13, 2024). Sanctioned party fined $10,000 for fabricated AI citations.
- Insurance Information Institute. “Expense Composition” Table (Archive). Figures represent industry aggregates and vary by line and jurisdiction.
- National Association of Insurance Commissioners. P&C and Title Mid-Year Industry Report 2024.
- S&P Global Market Intelligence. “U.S. Homeowners Insurers Combined Ratio Improves but Remains Above 100.” June 2024.
- Swiss Re Institute. Social Inflation: Litigation Costs Drive Claims Inflation (sigma 4/2024). September 2024.
- Milliman. “The Importance of Litigation Management for Insurance Carriers.” October 2024.
- McKinsey & Company. “The Future of AI in the Insurance Industry.”. Accessed October 2025.
