AI Didn't Replace Lawyers. It Gave Us the 48-Page Contract.

In 1972, two lawyers a few years out of school opened a storefront office in Van Nuys and called it the Legal Clinic of Jacoby and Meyers. Leonard Jacoby and Stephen Meyers were going after the large group of people who earned too much to qualify for legal aid and too little to retain a traditional law firm. To reach them, they cut costs wherever they could. They relied heavily on paralegal assistants and part-time specialists, charged a flat $15 for an initial consultation, skipped the traditional office building, and concentrated on simple but recurring legal problems that could be handled through standardized techniques.

The State Bar tried to suspend them for it. Not for the clinic itself, but for interviews they gave reporters about it. The Disciplinary Board recommended a 45-day suspension on the theory that talking to newspapers and a radio station about a low-cost legal clinic amounted to improper solicitation. The California Supreme Court reversed in Jacoby v. State Bar, 19 Cal.3d 359 (1977), holding that the rules could not constitutionally prohibit lawyers from cooperating in legitimate news coverage merely because the coverage might also help their business. The case is worth reading.

Every development that threatens to change how legal services are delivered seems to produce the same prediction: lawyers are about to become unnecessary. I have heard versions of it for most of my career.

I never thought the legal-clinic model would do what either its admirers or its opponents expected. Standardization was going to take over some work at the bottom of the legal pyramid, especially work that consisted largely of filling out forms and processing predictable matters. It was not going to take over the work of figuring out what a client actually needed.

Then came the internet and Google. Legal information that once required a trip to a law library became available to anyone with a browser. That was a huge improvement. It also produced clients armed with cases from the wrong jurisdiction, obsolete forms, and complete confidence that three hours online had resolved a difficult legal problem. Finding information had become easy. Knowing which information mattered remained difficult.

Now we have AI, and the prediction is back on a much larger scale. I regularly hear that AI will eliminate the need for lawyers in most matters.

I use AI every day. One of my favorite uses is to give it a complaint and ask what causes of action I may have missed. It might give me six. I expect several to be obvious losers, but if one deserves serious research, I have gained something. I also use AI to pressure-test arguments and review drafts. Our firm has written before about using AI to improve the practice of law, and I think lawyers who refuse to learn how to use it are making a mistake.

AI will eliminate legal tasks. It already has. What I am seeing so far, though, is something much less tidy: in some areas AI is creating more work for lawyers because it has made producing legal material almost effortless while leaving the hard part, judgment, largely untouched.

AI Made Bad Contract Drafting Scalable

Before generative AI, most non-lawyers simply couldn't produce a 25-page commercial contract. They might modify an old agreement, download a form, or write a short document covering the business terms that seemed important.

Those agreements often had serious holes, but at least the holes were visible. If the contract said nothing about an issue, everyone could see that, and the governing law would often supply an answer.

AI changed the economics of bad drafting. Anyone can now produce a polished commercial agreement in minutes. Ask what is missing and AI adds provisions. Ask it to make the contract safer and it adds more. A few prompts later, the five-page deal has become 25 or 35 pages of professional-looking legal prose.

The problem isn't getting enough clauses anymore. It is deciding which clauses belong, making them work together, and recognizing when one provision quietly contradicts another.

This is not a scientific study, but after reviewing well over a hundred contracts of this sort in the last year, I am seeing far more agreements so internally inconsistent that I cannot confidently say what they mean. Before generative AI, I might have encountered one such agreement a year.

We recently represented a German company in a dispute with an American company. The American company had drafted a 48-page contract that probably needed five to seven pages.

I needed to answer one basic question under that contract. Doing so required tracing six provisions scattered through the document. One incorporated another. The second contained an exception that sent me elsewhere. Another section appeared to undo the exception but depended on a definition found in a different part of the agreement. After going through the chain several times, I genuinely could not tell whether the contract meant A or not-A.

That uncertainty favored our client because the other side had drafted the agreement. Instead of debating one clear obligation, we could show the American company how difficult it would be to convince a court that its preferred reading was the right one. We sent a 20-page letter walking through the contradictions and litigation risks, and the matter settled for a relatively small amount.

Those 48 pages saved nobody time. They moved the work from drafting the deal to lawyers trying to determine what the deal meant after the relationship had soured.

How Overdrafting Creates Legal Work

More words don't necessarily prevent disputes. Sometimes they give both parties more reasons to believe they should win.

When a contract clearly favors one party, the weaker side has a reason to settle. When a short contract is silent, experienced lawyers can often look to the governing law and make a reasonable assessment of what a court is likely to do. A long agreement containing provisions that point in different directions changes that calculus. Each lawyer can find language supporting the client's position, and each party can persuade itself that its interpretation is the better one.

Uncertainty becomes optimism, and two optimistic litigants are harder to settle than one party that knows it has a bad case.

I cannot prove that AI-generated contracts are a meaningful reason good litigators seem so busy today. Too many other things affect litigation volume. But I am hardly alone in seeing AI create cleanup work. In an April 11, 2026 Financial Times article, Elizabeth Bratton reported on clients sending lawyers barrages of AI-generated emails, letters, questions, and other material that required additional review. Some firms were reconsidering fixed-fee arrangements because of the extra work. The article is here.

The broader legal market isn't behaving like a profession on the verge of extinction either. The Thomson Reuters Institute's Q1 2026 Law Firm Financial Index reported that demand grew 2.7 percent in the first quarter, nearly three times the industry's long-run average. That says nothing about whether AI caused the increase, but it should make us skeptical about predictions that lawyers are about to become unnecessary. Thomson Reuters' report is here.

What the China Contract Revealed

I wrote about a narrower version of this problem in June, when I explained why AI-generated China manufacturing contracts often look much better than they actually are. What I had not focused on then was the larger point: the same technology that makes contract language almost free can increase the value of the judgment needed to decide what that language should say.

A prospective client recently gave me an unusually good example. The company was preparing an agreement for a Chinese manufacturer that would help develop and manufacture a consumer projection night light. I have anonymized the companies, but the substance of the contract is real.

It was a surprisingly good-looking document. It dealt with tooling, project work product, and subcontractors. It required confidential material to be returned or destroyed. It even recognized the importance of the Chinese company chop.

This was not the sort of amateur agreement where the problems jump off the first page. The important problems were underneath.

The manufacturer's official Chinese legal name had not been confirmed, and English controlled the contract. Its remedies depended heavily on U.S.-style injunctive relief. Governing law and the dispute forum were still unresolved. Those issues went directly to whether the agreement would be useful if something actually went wrong.

At the same time, the contract contained provisions any sensible factory would question. One effectively required the buyer to approve the mold makers and component suppliers needed to build the product. Another could be read to restrict products the factory had already been making. A third barred the manufacturer from challenging any of the buyer's intellectual property anywhere in the world, apparently forever.

That produced a peculiar result. A competent Chinese lawyer reviewing the agreement for the factory could see that many of the provisions supposedly protecting the American buyer had little practical bite and might recommend signing it. Factory management could see immediately that some of the operational provisions were intolerable and might reject the agreement before a lawyer ever saw it.

The contract had become difficult to sign without becoming particularly useful to enforce.

A capable domestic lawyer can make this problem worse, even with good intentions. Home-state law feels comfortable. English feels like the obvious controlling language. Broad IP restrictions appear safer than narrow ones. Those instincts can be perfectly reasonable in a domestic transaction and badly wrong when applied to a Chinese manufacturer.

Contracts aren't graded on a curve. If the agreement binds the wrong entity, sends the dispute to the wrong place, or gives you remedies that don't work, getting twenty other clauses right does not rescue it. We have seen exactly what can happen when a China contract looks protective until something actually goes wrong.

I told this client not to send the agreement to the factory and not to pay us to repair it clause by clause. Fixing the document before figuring out the problem would have repeated the same mistake that produced it. First I needed to know what the manufacturer had already received and how deeply it would be involved in developing the product. Only then could we determine what the client still needed to protect and which agreements would do that.

The client had come to us asking how to fix an NNN Agreement. The more important question was whether an NNN Agreement was even the right agreement at that stage of the relationship.

The Hard Part Was Never the Words

That is the connection between the 48-page contract and the China contract. AI has automated much of the inexpensive part of contract drafting, which is generating language. The work that follows requires understanding the deal well enough to decide what the contract should actually do.

Our China manufacturing lawyers have hundreds of provisions accumulated and refined over years. Having those provisions doesn't draft the agreement for us. We first need to understand what the factory will control, what the parties will create together, and what the client needs if the relationship falls apart.

That is why the questions we ask before drafting a China manufacturing contract matter so much. The answers determine what belongs in the agreement. A template, whether written ten years ago or generated ten seconds ago, cannot make those decisions without somebody supplying the right facts and understanding why they matter.

The projection-night-light client is a good example. Because the manufacturer would participate in development and prototyping, the project raised questions about ownership of design changes, tooling, and the client's ability to move the finished product elsewhere. Those are development issues, not simply confidentiality issues, and they often call for a China Product Development Agreement rather than a longer NNN.

The question worth asking is not, "What clauses can we add?" It is, "What can actually go wrong here?"

Contracts Cannot Replace Verification

Even the right agreement depends on facts outside the document. A manufacturing professional commenting on my earlier LinkedIn post made a point I wish more companies understood: tooling and subcontractor provisions can be beautifully drafted while accomplishing very little because nobody checked what was actually happening at the factory.

Bills of materials change. Production gets moved. The factory named in one document may not be the company performing the work. If nobody knows what is happening operationally, the contract is protecting a version of reality that may no longer exist.

A clause saying you own a mold is useful, but you still want to know where the mold is. A prohibition against component substitution matters only if someone has identified the approved components. Before either provision matters, you need to know that the company signing the agreement is actually able to perform what it promises.

Drafting and verification are different disciplines. This is why overseas manufacturing due diligence belongs beside contract drafting rather than after it.

A Chinese supplier made a related point in the same LinkedIn discussion. Foreign customers sometimes spend weeks negotiating legal provisions no factory will accept while neglecting specifications, delivery requirements, and payment terms. Clients know those commercial terms better than their lawyers do, and good drafting starts there.

What AI Is Actually Replacing

The latest Bureau of Labor Statistics projections make an interesting distinction. BLS expects employment of lawyers to grow 5 percent from 2025 through 2035, faster than the average for all occupations, with about 28,700 openings each year. It acknowledges that some routine legal work will be automated but says the resulting efficiency is not expected to reduce overall demand for lawyers. Instead, it expects automation to let lawyers spend more time on other parts of their work, such as meeting with clients. The BLS lawyer outlook is here.

Now move one rung down the ladder. In the same August 27, 2026 update, BLS projects little or no employment growth for paralegals and legal assistants through 2035. It specifically says advances in technology, including AI, are expected to limit demand by making paralegals more efficient at research and document preparation. The BLS paralegal outlook is here.

That distinction has held since Van Nuys. Technology bears down hardest on work that can be reduced to a procedure. It does much less to the work of deciding which procedure applies, or whether any of them do.

There is a longer-term complication. Junior lawyers traditionally acquired judgment by doing research, reviewing documents, and preparing drafts that more experienced lawyers corrected. Brooke Loesby addresses this in The Hidden Shortage the Legal Profession Is Creating, in the July/August 2026 issue of the ABA's Law Practice Magazine. If AI removes too much of the work through which young lawyers learn, firms will need a different way to train the lawyers whose judgment clients eventually need. The ABA article is here.

That deserves its own discussion. It does not change what is happening now.

The Prediction Keeps Coming Back

I have heard the lawyer-extinction prediction for most of my career. Legal clinics were going to routinize law. Google was going to make legal knowledge free. Now AI can produce in seconds documents that once took a lawyer hours. Each development changed the profession, and AI will change it more than the other two.

What I see today is not the disappearance of legal work. I see contracts that are longer than they need to be and harder to interpret. I see sophisticated provisions drafted for the wrong country. I see clients arriving with far more legal material than they could ever have produced before, some of it useful and some of it badly confused. Somebody still has to decide what matters.

AI will keep getting better, and good lawyers should use it aggressively. But when it comes to contracts, I would rather have a client explain the actual deal to me on a napkin than hand me 30 pages of polished legal prose that nobody fully understands.