Alan Rozenshtein
banner
alanrozenshtein.com
Alan Rozenshtein
@alanrozenshtein.com
Member of technical staff @anthropic.com. Law professor (on leave) @umnlawschool.bsky.social. Contributing editor @lawfaremedia.org. Former DOJ. alanrozenshtein.com
A big part of my excitement at this opportunity is based on how how much support I've gotten from my institution and my colleagues (especially our fabulous dean @wmcgeveran.bsky.social). I'm very lucky!
@alanrozenshtein.com, associate professor of law, has taken a leave of absence from the University of Minnesota Law School to join @anthropic.com’s Rule of Law team as a member of its technical staff within the Anthropic Institute.
z.umn.edu/bhna
August 18, 2026 at 6:46 PM
Personal update: I've taken leave from @umnlawschool.bsky.social to join @anthropic.com as a member of technical staff researching AI and the rule of law at the Anthropic Institute. Excited to work with @botvinick.bsky.social and the rest of the team!
August 18, 2026 at 12:51 PM
Why I Quit the Tenure Track
The insufferable hypocrisy of elite academia
www.theatlantic.com
August 11, 2026 at 8:35 PM
In Chatrie, the Court did what it should have done in Carpenter: find that there was a Fourth Amendment search but be flexible on what counts as valid predication. papers.ssrn.com/sol3/papers....
Fourth Amendment Reasonableness After Carpenter
In Carpenter v. United States, the Supreme Court held not only that the Fourth Amendment applies when the government collects certain categories of third-party
papers.ssrn.com
June 29, 2026 at 2:59 PM
@robertwrighter.bsky.social has written a terrific new book about AI. I got to talk with him about it on @scalinglaws.bsky.social. podcasts.apple.com/us/podcast/t...
"The God Test": AI as Cosmic Reckoning, with Robert Wright
Podcast Episode · Scaling Laws · June 23 · 53m
podcasts.apple.com
June 23, 2026 at 11:47 PM
Turns out the U.S. did achieve unconditional surrender in the war in Iran...it was just our own surrender, not Iran's. www.nytimes.com/2026/06/17/u...
A Look at the Text of the Agreement Between the United States and Iran
www.nytimes.com
June 17, 2026 at 8:33 PM
Reposted by Alan Rozenshtein
"Even if Anthropic persuades the government to lift the control, the deeper question will remain: whether there should be, and if so under what circumstances, a licensing regime for frontier AI," writes @alanrozenshtein.com.
A Kill Switch for Frontier AI
The government is using export control law to force Anthropic to cut access to its most powerful models. The legal authority is plausible but the facts remain murky.
www.lawfaremedia.org
June 16, 2026 at 7:45 PM
Reposted by Alan Rozenshtein
On Friday, Anthropic received a letter from the Trump administration ordering the company to suspend all access to its Fable 5 and Mythos 5 AI models for foreign nationals. @alanrozenshtein.com explains the likely legal authority behind these export controls.
A Kill Switch for Frontier AI
The government is using export control law to force Anthropic to cut access to its most powerful models. The legal authority is plausible but the facts remain murky.
www.lawfaremedia.org
June 15, 2026 at 2:34 PM
If the USG’s export controls on Fable continue, it might set up the first big First Amendment case of the AI era.
June 13, 2026 at 3:13 AM
I'm tired of winning.
It seems overnight the White House has been replaced by a Monster Energy Drink branded UFC cage
June 10, 2026 at 1:11 PM
Excellent line: "Edison said that genius is 1% inspiration and 99% perspiration. But we see perspiration becoming increasingly automated." www.anthropic.com/institute/re...
When AI builds itself
Our progress toward recursive self-improvement, and its implications.
www.anthropic.com
June 5, 2026 at 12:21 AM
Reposted by Alan Rozenshtein
"Voluntary" isn't the floor of frontier AI policy in the Trump WH—it's the ceiling. And even that couldn't survive a few last-minute calls from industry. @kevintfrazier.bsky.social and @alanrozenshtein.com on what was killed in the AI EO and what comes next:

www.lawfaremedia.org/article/ai-g...
AI Governance by Phone Call
An executive order so deferential to the AI industry that it disclaimed any mandatory authority still couldn't survive a few last-minute calls.
www.lawfaremedia.org
May 26, 2026 at 10:35 PM
Reposted by Alan Rozenshtein
wow. Judge Henderson of DC Circuit (tho maybe her alone) is not at all buying DoD's "supply chain risk" designation of @anthropic.com . "For the life of me I don't see any evidence of maliciousness ... sabotage ... [or other words in definition].
To me this is just spectacular overreach by the dept"
May 19, 2026 at 2:00 PM
Reposted by Alan Rozenshtein
Recent reporting suggests that President Trump plans to sign an executive order promoting information sharing between AI labs and the government to identify risks by new models.

@kevintfrazier.bsky.social and @alanrozenshtein.com explain the “dominating AI by understanding AI” agenda.
Dominating AI Requires Understanding AI
AI dominance will require more than faster models—it will require breakthroughs in understanding, testing, and securing frontier AI.
www.lawfaremedia.org
May 13, 2026 at 2:15 PM
Congrats to @brendanballou.bsky.social for his new book! Excited to read.
May 12, 2026 at 4:52 PM
Reposted by Alan Rozenshtein
Almost a year ago, @lawfaremedia.org 's @alanrozenshtein.com and @benjaminwittes.lawfaremedia.org analyzed the nothing-burger 8647 seashell tweet and its worth revisiting:

www.lawfaremedia.org/article/the-...
The Situation: And What If He Meant It?
James Comey could have gone a lot stronger than “8647” and still not risked jail.
www.lawfaremedia.org
April 28, 2026 at 8:27 PM
Reposted by Alan Rozenshtein
Today James Comey was indicted by DOJ for the second time, this time over a social media post of a picture of seashells arranged to say “8647."

When it was first posted, @benjaminwittes.lawfaremedia.org and @alanrozenshtein.com explained why the mollusks in question don’t make out a threat.
The Situation: And What If He Meant It?
James Comey could have gone a lot stronger than “8647” and still not risked jail.
lawfaremedia.org
April 28, 2026 at 8:49 PM
Reposted by Alan Rozenshtein
Today, we’re launching a 5-week effort to sustain Lawfare’s independent analysis.

Lawfare is built on one idea: reliable, expert analysis delivered at the pace of news—without a paywall.

That only works because of our supporters. Become a $10/month supporter at lawfaremedia.org/support.
April 29, 2026 at 4:06 PM
Reposted by Alan Rozenshtein
New episode: @alanrozenshtein.com talks to @jassipannu.bsky.social and Doni Bloomfield about why governing biological data may be the most urgent AI risk policymakers aren't paying enough attention to. podcasts.apple.com/us/podcast/s...
Why Data Governance Is the Key to AI Biosecurity, with Jassi Pannu and Doni Bloomfield
Podcast Episode · Scaling Laws · March 24 · 50m
podcasts.apple.com
March 24, 2026 at 10:58 AM
Oh wow today's Anthropic hearing is going to be brutal for the government. storage.courtlistener.com/recap/gov.us...
March 24, 2026 at 10:26 AM