The explosion of legal AI creates vast opportunities, but contrary to the views of Box CEO Aaron Levie, this won’t mean more work for law firms.
Generative AI has driven an explosion of possibilities and given non-experts access to fields that were previously out of reach. Now, almost anyone can be a graphic designer, a software developer or a copywriter and the output is good enough, and affordable, for the average person. LinkedIn feeds are now flooded with posts from out-of-work professionals struggling to find work and bemoaning the fact that AI just isn’t as good as the work of a human.
The fears of AI taking jobs also apply to the legal professional. However, when it comes to legal AI, things aren’t always that simple. Specialised legal AI tools have brought about radical cost and efficiency savings for legal and non-legal teams alike, however they also generate work: using legal AI, more contracts can be drafted and analysed, more queries can be answered. This is Jevons Paradox at work: as technology drives efficiency and costs drop, its usage increases exponentially.
Does generative AI mean more lawyers?
This explosion in AI-generated legal content, argues Aaron Levie, Co-Founder and CEO of enterprise cloud company Box, is creating more – not less – work for lawyers. Levie argues that AI has done nothing to clear the bottleneck in the court system and more qualified lawyers will be needed to adjudicate the dramatic increase in legal disputes.
As he outlined in a post on social media platform X, more lawyers are needed in the same way that the advent of personal computers and the internet drove an increase in legal work. He cites figures from the American Bar Association (ABA), that chart this trajectory, from nearly 400,000 attorneys in the United States in 1975 to more than 1.37 million at the end of 2025.
Levie argues this trajectory will continue. Based on what I have seen from the early stages of implementing legal AI solutions, however, this is not the pattern I see playing out. I agree that AI will expand the market for legal work, but it won’t bring more work for lawyers. Instead, I believe legal AI will drive a shift away from law firms and companies will choose to do more of this work in-house.
Specialized legal AI will drive work from law firms into companies
Levie’s argument assumes that the demand for legal services will derive from accessible and generalized tools, such as Claude or Copilot. However, on the other side of the fence are specialized legal AI solutions that are dramatically reducing time spent on legal tasks like reviewing contracts. This AI specifically targets workflows at law firms and in-house legal teams, meaning that masses of manual work have now been eradicated. Specialized legal AI replicates and automates much of the repetitive legal work that has been the bulk of a lawyers’ day-to-day responsibilities. And for years, corporates have accepted that they had to pay high rates for overskilled professionals to do this manual work.
Traditionally, the largest law firms recruited graduates to scan documents, review cases, and compile summaries and arguments. Work that took days to produce can now be done in minutes, seconds even. With competent summaries being produced in a fraction of the time, for a fraction of the cost, law firms are having to rethink their business model of charging by the hour.
With AI tools that replicate and automate specific workflows, the need to seek outside counsel will reduce as companies will be able to do the bulk of the legal work themselves.
Legal AI empowers in-house and business users
Levie’s scenario focuses on the courts, but I see a greater opportunity outside of this system. I believe that automation and AI tools that are trained for specific use cases will empower legal teams and business users within companies. There may be an expansion in the amount of legal content, and work for professionals, but this will be offset by the automation AI brings. For in-house teams, they will likely be just as busy as they always have been.
Take the example of insurance companies and claims handling. The bulk of this work, which includes reviewing and cross-checking legal documents, can be automated, and AI can check, for example, whether a policy is valid and covers the claim. It can also construct a timeline and spot where details are missing and generate a message to request additional information. A human in the loop remains necessary to oversee the process, but there will not be the bottleneck that Levie describes with the court system. Agreed, there will be an expansion of the market for legal work, but the bulk of that work won’t be done by law firms; it will be done by in-house teams and business users who have been empowered by the use of specialized legal AI.
Join the conversation!