The Justice Gap Paradox: AI Will Create More Legal Work, Until It Doesn't

Jensen Huang, the CEO of Nvidia, recently pushed back against the idea that AI would trigger a "white-collar apocalypse." He argued that increased productivity allows us to build more of the things we need. The legal profession is a perfect case study. There are plenty of lawyers, and plenty of demand for them. The problem is far too few people can afford them.

That mismatch is called the justice gap, and it is usually presented with somber statistics. We will get to those. But here is the twist: AI might close the gap by giving lawyers more work than ever … until it doesn't.

The scale is staggering: a 2022 justice gap study found low-income Americans received no or insufficient legal help for 92 percent of their substantial civil legal problems, and a 2024 California study showed only 18 percent of residents sought help for any civil legal problem.

When people do not pursue a legal problem, it escalates and returns as an emergency. Evictions become homelessness. Benefit denials become health crises. Family disputes become years of instability. The justice gap is measured not only in unmet needs, but in the downstream costs of inaction.

The Return of the Small Claim

Small claims filings have been shrinking for decades. In New York, they fell from more than 90,000 in 1997 to about 22,000 in 2021. Across 32 states, filings dropped 32 percent between 2018 and 2022. Disputes simply became too costly and inconvenient to pursue.

AI changes that equation. By lowering the time and effort required, small disputes are worth filing again, especially for people representing themselves. Tools that draft documents, prepare evidence, and guide litigants through court procedures give pro se parties the confidence to pursue cases they once abandoned. This could spark a resurgence in small claims, as everyday disputes that seemed impractical finally make it into court.

Legal Aid at Higher Stakes

Legal aid organizations face a different challenge. Their clients already bring significant matters involving housing, benefits, immigration, or family law. The barrier here is that the staff doesn’t have the capacity to take them all on.

Here, AI helps in another way. Intake and triage tools can process more applications with fewer bottlenecks. Document assembly can cut hours of work into minutes. Lawyers who once had to decline half of all requests for help can open more files, not by working longer, but by working faster. Each new case then triggers further legal activity, such as a defended eviction leading to a repair claim or a successful benefits approval creating new compliance questions.

More Plaintiffs, More Defendants

In a constrained market like legal services, AI-driven efficiency expands access. It also increases work for the better-resourced. This can be thought of as the justice gap dividend.

As more claims enter the system, the other side follows. Tenants who bring actions against landlords require responses. Workers who file wage or benefits disputes create matters for employment counsel. Small businesses that rarely faced litigation may now see a steady stream of claims.

A Temporary Surge?

Some experts, like Dario Amodei of Anthropic, have warned that AI could threaten a large share of entry-level white-collar jobs. If a system can reason about the law and produce advice that is indistinguishable from that of a human lawyer, the dividend changes shape. People may not need more lawyers; they may need legal systems that allow AI to serve them directly.

When the automobile first arrived, demand briefly spiked for stronger horseshoes, better carriages, and improved buggy parts. For a moment, it looked like the horse-and-buggy industry was being revitalized. A generation later, it was gone.

AI may follow a similar arc. By lowering costs and expanding access, it creates both more cases and more lawyers in the short term. But over time, the very tools that revived these claims could also reduce the need for lawyers to handle them.

Garfield.Law and the Next Phase

This shift is already underway. In 2025, Garfield.Law in the United Kingdom gained approval to run a debt recovery practice almost entirely through AI. The firm charges £2 per letter, automating what paralegals once did. The justice gap narrowed because the AI itself did the work, not because more lawyers were available.

Debt recovery may be the first area where this pattern is visible, but it will not be the last. Routine disputes, standardized claims, and document-heavy matters are especially vulnerable to automation. The same systems that expand access today may reduce the role of lawyers tomorrow.

The Arc of the Paradox

AI is already changing who can pursue claims, how quickly they move, and how many people gain access to justice. In the short term, that means more small claims, more legal aid clients being served, more pro se litigants and more defense work for businesses.

In the long term, it raises a sharper question: will the justice gap be closed by adding more lawyers, or by making lawyers unnecessary for routine cases?

For now, AI is expanding the reach of justice and preventing problems from spiraling into emergencies. The paradox is that in solving these problems, it may eventually close the gap and close the door on parts of the profession itself.