David Kemp

Anthropic Claude status showing downtime five of the past six days.

Anthropic is having a rough few days.

This is beautiful. Added to my current playlist along with the Chicks’ March March.

Fully expecting to see an E.O. purporting to replace MLK Day with DJT Day 🤮

Supreme Court holds that a Bivens claim requires both the victim and the federal agent to be named Bivens.

Judges’ AI Blunders Spark Debate on Technology Use in Courts When questioned by the chairman of the Senate Judiciary Committee over a flawed June opinion, a federal judge blamed an intern’s use of generative AI for the error and banned use of the technology in ...

“...a better approach is to teach students how to use it effectively and when not to use it.”

#RutgersLaw Adjunct Professor @davidkemp.bsky.social on AI bans in the legal realm: https://news.bloomberglaw.com/business-and-practice/judges-ai-blunders-spark-debate-on-technology-use-in-courts
#lawsky

Your Students Need an AI-Aware Professor Here’s a sustainable plan to bring you up to speed on a technology that academe can’t afford to ignore.

In Advice: A quiet disaster is unfolding in college classrooms around this country as AI becomes an undeniable reality for students — with little or no guidance from professors. Here’s how to get up to speed on this technology. chroni.cl/4j5urjK

Legal practice requires careful adherence to procedural rules. In the United States, few are more complex than those found in The Bluebook: A Uniform System of Citation. Compliance with this system’s 500+ pages of byzantine formatting instructions is the raison d’être of thousands of student law review editors and the bête noire of lawyers everywhere. To evaluate whether large language models (LLMs) are able to adhere to the procedures of such a complicated system, we construct an original dataset of 866 Bluebook tasks and test flagship LLMs from OpenAI, Anthropic, Google, Meta, and DeepSeek. We show (1) that these models produce fully compliant Bluebook citations only 69%-74% of the time and (2) that in-context learning on the Bluebook’s underlying system of rules raises accuracy only to 77%. These results caution against using offthe-shelf LLMs to automate aspects of the law where fidelity to procedure is paramount.1

New paper:
Matthew Dahl, Bye-bye, Bluebook? Automating Legal Procedure with Large Language Models
arxiv.org/pdf/2505.02763
#LegalWriting
#Citation
#LegalTech

Fat and furious

Congratulations!

Fighting the Last (Trade) War: Trump Ignores the Coming AI Revolution Cornell Law professor Michael C. Dorf critiques the Trump administration’s tariff policies and broader economic strategy, arguing that they are misguided in the face of rapidly advancing technology, p...

My latest column for @justiaverdict.bsky.social argues that even if Trump's tariffs boost U.S. manufacturing (a very big "if"), they won't create many jobs. His effort to restore a 19th century economy leaves us woefully unprepared for the disruptive force of AI in the 21st century. 👇

When Artificial Intelligence Workshops Your Article For You—And It's Really Useful So this is cool.

"When Artificial Intelligence Workshops Your Article For You—And It's Really Useful" My first post at a new blog about a new app. cc: @christianturner.bsky.social
blog.dividedargument.com/p/when-artif...

Pitch it to the editor of @justiaverdict.bsky.social

@judgefergusontx.bsky.social

So they told us what they were going to do to ALL of us. Here is a tracker on their progress in destroying our country:
www.project2025.observer

This article makes some good points, but I think it fails to consider that, based on how the “reasoning” models work, they can be trained to be not just “correct” but “good,” if the developers so choose.

Rather than pitting your writing against Claude’s, imagine all you could produce with Claude as an assistant.

Something like this…

No real consensus, but Claude is very good, and soon ChatGPT’s newest models will soon be available for free.

For law students who wonder why 1L spends so much time on the holding/dicta distinction: The rule of law itself sometimes hangs in the balance.

Phenomenal.

The ABA supports the rule of law. Read full message: www.americanbar.org/news/abanews...

I celebrate this annually! I’ll never forget how my amazing Berkeley Law students all came to our Zoom class the next day as cats.

Dear journalists: Now, more than ever,

✨The lawyers need to see the actual court filings.✨

Yes, I know you summarized. Yes, I know your non-attorney readers don't care.

But we need to see the documents. The actual documents.

If you're pulling docs from PACER, use RECAP & give us a link. Please.

Using novel data on financial disclosures of judges and the civil case information of public firms across district courts in the United States between 2000 and 2021, we document multiple instances in which judges fail to recuse themselves in cases where they hold common stock. We find robust evidence that judges appear to rule more favorably and take longer when deciding over conflicted cases than when deciding over unconflicted ones. We also find that investors are positively surprised by these outcomes, as reflected by stock returns to case outcomes. Collectively, our results provide the first large-scale evidence of the relation between judges’ financial holdings and case outcomes.

New paper using our data shows a correlation between case outcomes and judicial investment conflicts. "We find robust evidence that judges appear to rule more favorably and take longer when deciding over conflicted cases" papers.ssrn.com/sol3/papers....