AI Model GPT-3 Says Legaltech is “Full of Hype”: Artificial Intelligence Trends

I’ve been starting to hear quite a bit about GPT-3, which is touted as the most powerful language-generating Artificial Intelligence (AI) tool to date.  It’s a transformer-based language model released by OpenAI with 175 billion parameters.  GPT-3 even writes blog posts, and a college kid just used it to reach the number one spot on Hacker News!  An AI engine that writes blog posts?!?  Ruh-roh!  Should I be worried?  Well, wait until you see what GPT-3 had to say about “Legaltech”.

Thanks to Mike McBride’s weekly online newsletter (latest edition here), I found this article from Genie AI, which, after a brief introduction, is a blog post written by GPT-3 about “Legaltech”.  Here are a few “observations” from GPT-3 – bold/italic text added by the Genie AI team – with my own comments in blue in between.  See what you think.

The legaltech industry is in the process of transforming the way we interact with the legal system, and the growth and evolution of the legaltech market has been remarkable. The current growth is exciting and makes for a very exciting future.

The current state of legaltech is strong and getting stronger. As more and more companies continue to explore and develop legaltech tools and solutions, the more we’ll see these solutions come to fruition and have a lasting impact on the industry.

So far, so good.  Makes sense to me.

The biggest problem for the legaltech industry is that it has not really produced anything of use to society, or at least not anything of real value. The tech sector is full of stories of big failures, which can be explained by the excessive amount of funding that has been poured into it.

Wow, GPT-3 is harsh!  No filter there!  Looks like all of us in “legaltech” have been working for years and years – and have “not really produced anything of use to society”.  I feel like such a failure.


The legaltech sector is not really different. It is full of hype, with stories of billion dollar valuations for companies that have not yet made a cent in profit. It is now even possible to find companies in the sector that have not even launched yet.

I won’t argue about “legaltech” being “full of hype” – we see a lot of hype in this industry.  But, “stories of billion dollar valuations for companies that have not yet made a cent in profit”?  What are we, the next bubble?

One problem is that the legaltech industry is dominated by entrepreneurs who have little understanding of the legal sector. This is a problem because the legal sector is one of the most difficult sectors in which to operate.

I agree the legal sector is very difficult.  Is the industry being “dominated by entrepreneurs who have little understanding of the legal sector”?  Hmmm…what do you think?

The one hope for the legaltech industry would be that if you, the lawyer, are no longer relevant, there is always the judge. They have a life-tenured job and can’t be fired. If the industry can get enough AI and other technologies to make you obsolete, then the judge is still needed. So, the industry may still have a future in lobbying to keep judges.

WHAT?!?!  O! M! G!  Wow.  Sorry, all you lawyers out there, you’re no longer relevant unless you can become a judge.

To stop the hype in legaltech, and so that companies can actually start making money, legaltech companies should focus on automating tedious, mundane tasks that lawyers don’t want to do. Legaltech companies should do things that lawyers don’t want to do, and that they will pay for. This will be the only way to get the law firm business, and the only way to generate real, long-term, profitable growth.

Well, duh!  That’s where we went wrong!  Now, I understand – do “things that lawyers don’t want to do, and that they will pay for”.  Of course, it’s so obvious.

So, according to GPT-3, “legaltech” is “full of hype”, “has not really produced anything of use to society”, is “dominated by entrepreneurs who have little understanding of the legal sector” and needs to do “things that lawyers don’t want to do, and that they will pay for”.  Oh, and lawyers should plan to become judges before they’re obsolete.

As for that college student, Liam Porr, who reached the top of Hacker News with his GPT-3 generated post?  Apparently, the trick to generating content without the need for much editing was understanding GPT-3’s strengths and weaknesses. “It’s quite good at making pretty language, and it’s not very good at being logical and rational,” said Porr. So he picked a popular blog category that doesn’t require rigorous logic: productivity and self-help and added a catchy title to it – “Feeling unproductive? Maybe you should stop overthinking”.

Look, as I wrote last week, I think AI technology has tremendous capabilities to simplify the discovery process.  It’s already been doing that for years with machine learning to support predictive coding and other workflows and I think it will be a game changer in identifying Personally Identifiable Information (PII) for compliance purposes and potential redaction, using natural language processing (NLP) and linguistic models to help in that effort.  But, when it comes to writing what it “thinks”?  Well, GPT-3, “Maybe you should stop overthinking”.  At least for now.

BTW, click on the GPT-3 demo video in the Genie AI article to see what GPT-3 thinks about “legaltech” pricing! 😉

So, what do you think?  Have you implemented AI technology in your eDiscovery workflows?  If so, where?  Please share any comments you might have or if you’d like to know more about a particular topic.

Disclaimer: The views represented herein are exclusively the views of the author, and do not necessarily represent the views held by my employer, my partners or my clients. eDiscovery Today is made available solely for educational purposes to provide general information about general eDiscovery principles and not to provide specific legal advice applicable to any particular circumstance. eDiscovery Today should not be used as a substitute for competent legal advice from a lawyer you have retained and who has agreed to represent you.


  1. You nailed it again, Doug. I think it proves a point I’ve been trying to make for years: There’s no real AI in legal (yet).

  2. Thanks, Mike. Actually, I feel there is some real AI in legal, but, in many cases, people expect it to be a lot more than it is — at this stage at least.

  3. Hi Frank, thanks for the comment. I based my reporting on 750 billion parameters on this article (, but I see elsewhere that it’s being reported as 175 billion (including MIT’s Technology Review here:, so that must be the correct number. Who am I to argue with MIT? 😉 I’ve updated the post.

Leave a Reply