Anna Guo reports on a study comparing the performance of lawyers and AI apps. Key takeaways:
1. Several AI tools matched, and in some cases outperformed, lawyers in producing reliable first drafts.
2. The top AI (Gemini 2.5 Pro) marginally outperformed the top individual human lawyer: 73.3% vs. 70% reliability rate.
3. Specialized legal
There’s a significant debate on how soon computers may achieve artificial general intelligence. The AlphaZero Project–which led one of its leaders to receive the Nobel Prize for Chemistry–is an important data point in the discussion. Here’s the intro to my article at LLRX.com.
Will computers ever achieve the holy grail of artificial general intelligence
My review of the new book, The AI Con: How to Fight Big Tech’s Hype and Create the Future We Want (Harper 2025), by Emily M. Bender and Alex Hanna, is available at LLRX.com.
It’s the first negative book review I have written since I began writing reviews for The Federal Lawyer back in 1995. My
AI in the legal world is not a sentient being. It doesn’t argue cases, it doesn’t replace lawyers, and it certainly doesn’t understand the law in the way the best lawyers can.
What it can do is process large amounts of text quickly and find patterns in ways that mimic certain kinds of legal work—especially the repetitive
From my review of How to Think About AI: A Guide for the Perplexed at Attorney At Work:
With over 500 generative AI apps for lawyers cataloged by LegalTech Hub in March 2025, the proliferation of such tools shows no signs of slowing. In this environment, nuanced understanding is more critical than simply another
Too many books today are getting it all wrong about the risks of AI. Many vendors and dreamers are telling us there is no risk. It’s all going to be peachy and positive. Others are screeching that AI will inevitably lead to the Apocalypse.Richard Susskind gets it right. His new book, How To Think About
Richard Susskind’s new book How to Think About AI: A Guide for the Perplexed suggests three ways AI might change the workplace–and which one to worry about. He distinguishes three concepts:
Today’s chess-playing computers can crush the best human players without breaking a sweat. This wasn’t always true. A couple of decades ago teams of the strongest humans and the most powerful computers were stronger than either humans or computers alone. These teams were sometimes called “centaurs.” They combined the strength of a mighty beast
Richard Susskind’s new book How to Think About AI has a warning for professionals:
Professionals see much greater scope for AI in disciplines other than their own. Doctors are quick to suggest that AI has great potential in law, accounting and architecture, but instinctively they tend to resist its deployment
The new book AI Snake Oil calls out the major “hype superspreaders” fueling today’s AI bubble: