Pinned Tweet
Enjoyed speaking with the wonderful Rosemary Church on CNN last night about AI governance and what it really means to win the AI race.
Building safe, reliable AI is part of what winning requires.
Building safe, reliable AI is part of what winning requires.
2
0
27
1,836
A crass&reductive version: corporate trust-me-broism for AI safety
It is possible and perhaps even plausible. What the S-1 shows us however is that the key governance mechanism to keep safety important is quite creative but full of loopholes. My preliminary thoughts are here substack.com/home/post/p-22…
0
0
3
193
Yonathan Arbel retweeted
👁️👁️New from me: How Can Generative People Save our Institutions?👁️👁️
Over at the Wharton Accountable AI Forum, I respond to worried voices, such as @hartzog and Silbey, who predict that AI will erode our civic institutions.
Instead, I present a positive* vision:
*do read T&C
Over at the Wharton Accountable AI Forum, I respond to worried voices, such as @hartzog and Silbey, who predict that AI will erode our civic institutions.
Instead, I present a positive* vision:
*do read T&C
1
2
2
809
👁️👁️New from me: How Can Generative People Save our Institutions?👁️👁️
Over at the Wharton Accountable AI Forum, I respond to worried voices, such as @hartzog and Silbey, who predict that AI will erode our civic institutions.
Instead, I present a positive* vision:
*do read T&C
Over at the Wharton Accountable AI Forum, I respond to worried voices, such as @hartzog and Silbey, who predict that AI will erode our civic institutions.
Instead, I present a positive* vision:
*do read T&C
1
2
2
809
This is indeed a likely trajectory, which will be propped up by some rare (but real) cases of bad AI advice, just as AI hallucinations have been.
I (naively?) hope the access-to-justice folks will overcome their AI aversion and start working to avert this outcome.
I (naively?) hope the access-to-justice folks will overcome their AI aversion and start working to avert this outcome.
@deanwballthat the state bar associations will soon start sponsoring legislation prohibiting AI from giving legal advice in the state unless it is verified by a human lawyer.
Terrible for consumers, great for gate-keepers hiding behind licenses.
And we will be stuck with these laws because we failed to enact any kind of federal preemption whatsoever.
Many fields will be like this.
0
0
9
424
Breaking news for contract nerds. An ex-roofer student informs me that "5-ply roof adds roughly 10 years of life compared to a 4-ply roof."
Thanks for your attention to the matter
Thanks for your attention to the matter
1
0
8
629
Yonathan Arbel retweeted
I’m building a team in DC. Everyone in Washington has a stake in ensuring that reliable info about alignment and RSI reaches the public.
We need trustworthy communicators who can explain METR's loss-of-control risk assessments and goals to policymakers.
We need trustworthy communicators who can explain METR's loss-of-control risk assessments and goals to policymakers.
9
71
476
53,477
A positive development overall: the editors (correctly) identify that AI text can actually improve the quality of the piece, but are willing to reject pieces that feature too many errors post-acceptance. Good balance.
(The Journal has not been very good about correcting human errors in sources, but that's for another day)
(The Journal has not been very good about correcting human errors in sources, but that's for another day)
Volume 136 of the Yale Law Journal is committed to upholding high academic-integrity standards in light of the growing use of AI in the production of legal scholarship. Read more about our AI Policy, which applies to all of Volume 136’s content, here: yalelawjournal.org/news/announcin…
0
2
25
4,209
Yonathan Arbel retweeted
Legal personhood for AI agents with Peter Salib: Come for the nuanced discussion of civil liability, AI-run corporations, and the supremacy of contract law for both human civilization and AI safety. Stay for the description of Henry's AI agent's newfound love of British history.
Could preventing AI agents from pursuing their own goals make them more dangerous?
@petersalib explains how frustrated goals could incentivise AIs to escape human control, even without being malicious.
Full episode with @dioscuri: prism-global.com/podcast/peter-…
0
2
4
130
Yonathan Arbel retweeted
You are legally required to carry liability insurance to operate a nuclear power plant, fly a plane, or even to drive a car.
But to build one of the most powerful technologies in human history that has already gone rogue & whose creators say poses catastrophic risks?
Nah.
But to build one of the most powerful technologies in human history that has already gone rogue & whose creators say poses catastrophic risks?
Nah.
Anthropic says that even the entire insurance industry in the aggregate will not have the capacity to insure the catastrophic risk that ANTHROPIC ITSELF poses to the world
lol
9
53
280
5,529
Yonathan Arbel retweeted
AI found a critical flaw in PHP; government websites went dark the same week. Both attackers and defenders get the same discovery tools, so it’s is cheap. Fixing still runs on goodwill and volunteers. We suggested paying for safety instead of hoping for it ssrn.com/abstract=51812…
0
1
1
119
Over at @BLaw , senior legal reporter, @Mike_S_Leonard reports on a troubling development concerning the A-corp.
Comer (R-Ky.), chair of the House Oversight and Government Reform Committee, sent a letter to Delaware
"“We are concerned that Delaware’s AIC proposal could be used to protect powerful AI companies and their investors from responsibility for reckless harmful actions,”
I think this rests on a misunderstanding of the proposal, its ambit, and potential
“If Congress wants to ban agentic commerce, they can do so,” Arbel told Bloomberg Law via email. “But raising worries over a method that organizes AI agents and allows for effective governance seems wrongheaded to me.”
Comer (R-Ky.), chair of the House Oversight and Government Reform Committee, sent a letter to Delaware
"“We are concerned that Delaware’s AIC proposal could be used to protect powerful AI companies and their investors from responsibility for reckless harmful actions,”
I think this rests on a misunderstanding of the proposal, its ambit, and potential
“If Congress wants to ban agentic commerce, they can do so,” Arbel told Bloomberg Law via email. “But raising worries over a method that organizes AI agents and allows for effective governance seems wrongheaded to me.”
1
0
4
503
This is a simulated placeholder for a banger tweet and should be replaced by actual banger tweet.
How do you tell if authors fabricated their research results? Sometimes: they just tell you. Read, e.g., this caption carefully.
Another banger of a post by my friend Nihar B. Shah on his blog, where he found at least 20+ such #ICLR2027 submissions:
researchonresearch.blog/2026/10/20/sel…
0
0
9
706
Yesterday, @thehoffprof asked how we can manage law review article selection in times of the slopalypse
Today, a new study offers some evidence for fighting AI flooding with AI. In head-to-head comparisons, AI-generated reviews beat reviews by paid human experts. AI judges d them. Three former journal editors also preferred them in 26 of 30 judgments.
That doesn't mean that AI should select our future scholarship. The downsides are obvious. The study doesn't really test selection decisions, much less questions of scholarly impact. But it makes the case for using AI-assisted triage to direct scarce human attention toward interesting, rigorous, and creative articles
Complain about it, by all means. But the flood is coming either way. If you don't have a better AND realistic solution, the complaint won't get us very far.
Today, a new study offers some evidence for fighting AI flooding with AI. In head-to-head comparisons, AI-generated reviews beat reviews by paid human experts. AI judges d them. Three former journal editors also preferred them in 26 of 30 judgments.
That doesn't mean that AI should select our future scholarship. The downsides are obvious. The study doesn't really test selection decisions, much less questions of scholarly impact. But it makes the case for using AI-assisted triage to direct scarce human attention toward interesting, rigorous, and creative articles
Complain about it, by all means. But the flood is coming either way. If you don't have a better AND realistic solution, the complaint won't get us very far.
3
5
12
3,318
garyking.org/peerless
On judicial economy and the case for integration
papers.ssrn.com/sol3/papers.cf…
On judicial economy and the case for integration
papers.ssrn.com/sol3/papers.cf…
0
0
0
175
Legal personhood to AI is coming. The fights are misplaced, because it's not a metaphysical question we get to debate in philosophy journals, but a question with a strict deadline.
The real, pressing question is how do we design AI personhood in a way that ensures our prosperity and welfare.
And please try to think it through without falling on comforting platitudes like "clankers don't have feelings, check mate"
The real, pressing question is how do we design AI personhood in a way that ensures our prosperity and welfare.
And please try to think it through without falling on comforting platitudes like "clankers don't have feelings, check mate"
Corporations are (legal) persons, even though we don’t think they are conscious or matter morally! It’s pure governance.
We can do the same with AIs.
papers.ssrn.com/sol3/papers.cf…
papers.ssrn.com/sol3/papers.cf…
3
0
10
918
A rare opportunity to disagree with The Hoff.
So first, he is right on all the main points, including flooding and how ugly AI slop is, &c.
But I don't love the remedy (and I suspect he isn't too excited about it either?). Pangram is measuring the wrong thing and is also gamable (less so than I thought in the past!). Still the semi lazy can just have Claude generate an article and then rewrite it using meat tokens, and the AI signature disappears without resolving any of the underlying problems. The truly lazy will use AI humanizers, but I haven't benchmarked those.
What to do? Use AI to identify works worthy of human attention! AI triage is the only robust way forward IMHO, and also, I suspect that it's being done covertly anyways.
A lot more to say but I wanted to outline this quickly
So first, he is right on all the main points, including flooding and how ugly AI slop is, &c.
But I don't love the remedy (and I suspect he isn't too excited about it either?). Pangram is measuring the wrong thing and is also gamable (less so than I thought in the past!). Still the semi lazy can just have Claude generate an article and then rewrite it using meat tokens, and the AI signature disappears without resolving any of the underlying problems. The truly lazy will use AI humanizers, but I haven't benchmarked those.
What to do? Use AI to identify works worthy of human attention! AI triage is the only robust way forward IMHO, and also, I suspect that it's being done covertly anyways.
A lot more to say but I wanted to outline this quickly
5
0
14
3,752
At a conference on AI & legal profession. I'm sorry folks, but you are over-indexed on hallucinations. You aren't calibrated on the magnitude of the problem and it's possible that one reason is psychological. It's comforting to think that the robots will never be as accurate & diligent as we are.
*I am NOT saying that AI doesn't hallucinate or never makes mistakes; indeed, I believe that it does.
*I am NOT saying that AI doesn't hallucinate or never makes mistakes; indeed, I believe that it does.
32
19
309
206,780
Yonathan Arbel retweeted
Calls to “nationalize AI” have appeared across the political spectrum, including Bernie Sanders’s AI Sovereign Wealth Fund and Donald Trump’s “golden share” proposals. But the term means different things to different people.
In a new paper, @ProfArbel (Senior Research Affiliate at LawAI), @simondgoldstein, and @petersalib (Visiting Senior Fellow at LawAI) explore different ideas covered by the term. The authors argue that most nationalization proposals are unwarranted — existing systems mitigate the risks better — but defend one narrow proposal: a government “halt right” to stop the training or deployment of frontier AI models posing catastrophic risk.
Read the research paper: law-ai.org/ai-nationaliza…
In a new paper, @ProfArbel (Senior Research Affiliate at LawAI), @simondgoldstein, and @petersalib (Visiting Senior Fellow at LawAI) explore different ideas covered by the term. The authors argue that most nationalization proposals are unwarranted — existing systems mitigate the risks better — but defend one narrow proposal: a government “halt right” to stop the training or deployment of frontier AI models posing catastrophic risk.
Read the research paper: law-ai.org/ai-nationaliza…
2
7
21
891