Three and a half years ago, when ChatGPT v1 was released, many predicted that lawyers would be the first on the chopping block. The time is coming now, just not in the way most people expected.
Most people expected that capable lawyers would be replaced by capable AI. That’s not what’s happening.
As I’ve written previously, the AI tools created by large legal publishers like Westlaw and Lexis-Nexis are disappointing. The LLMs themselves, Claude and ChatGPT, are much better, but even they are less capable, in most instances, than a first-year lawyer. LLMs still hallucinate and will always hallucinate – they invent cases or statutes that don’t exist and pretend a statute says one thing when it says the opposite. High-profile law firms have been sanctioned recently for depending on AI.
And yet, with OpenAI, Anthropic, and SpaceX moving toward IPOs, the hype from the AI industry is a locomotive that can’t be stopped.
Many non-lawyers believe that AI has already reshaped the legal industry. Online, non-lawyers wonder cynically why lawyers are still expensive, given that AI is doing all the work. The General Counsel of a large company demanded that firms cut their rates to reflect AI efficiencies. That AI can draft legal documents as well as it writes code is taken as a given – if it hasn’t happened today, it will happen tomorrow. Caught in the frenzy, Kirkland and Ellis just announced plans to spend $500 million on AI, even though it will be obsolete long before it’s finished.
The apotheosis is the “AI-native law firm”, and the apotheosis of that is a firm created recently by a lawyer who used to work for OpenAI. He raised millions of dollars from big-name VCs. The firm uses AI, presumably trained on typical forms, to create drafts of contracts. Then the drafts are sent – I’m not making this up – to 50+ outside flesh-and-blood lawyers to review.
Is that an “AI-native law firm” or a normal 50-lawyer law firm of uncertain quality?
We could call it a “computer-native law firm” instead, with as much meaning. And back in the day, law firms were “typewriter-and-books-native law firms”.
Yet the locomotive moves on. Practicing lawyers know the shortcomings of AI, but many of their clients don’t. Consequently, lawyers are losing work to AI, as predicted three years ago, but the work produced by the AI is vastly inferior. It’s no exaggeration to say that lawyers are losing work not to the technology but to the hype.
In the short term, many businesses will use legal documents and make legal decisions that are flawed. But flaws in legal documents and legal decisions often remain hidden for a long time. Maybe, with the passage of time, the AI-for-law tools will live up to the hype. Maybe, after the IPOs, the hype will subside and the market will adjust. Meanwhile, it’s going to be difficult for a lot of law firms and a lot of law school graduates, squeezed between the hype and the reality of AI.
To take my mind off it, I’m traveling to an AI-native pond to catch AI-native fish.
Questions? Let me know.
Markley S. Roderick
Lex Nova Law
10 East Stow Road, Suite 250, Marlton, NJ 08053
P: 856.382.8402 | E: mroderick@lexnovalaw.com