Sen. Bernie Sanders (I-VT) and Rep. Greg Casar (D-TX) introduce legislation to ban Artificial Superintelligence and temporarily pause advanced AI development

Matrice Jacobine

Sen. Bernie Sanders (I-VT) and Rep. Greg Casar (D-TX) introduce legislation to ban Artificial Superintelligence and temporarily pause advanced AI development — LessWrong

225 Sen. Bernie Sanders (I-VT) and Rep. Greg Casar (D-TX) introduce legislation to ban Artificial Superintelligence and temporarily pause advanced AI development

by Matrice Jacobine

3rd Sep 2026

2 min read

225

This is a linkpost for https://www.sanders.senate.gov/press-releases/news-sanders-casar-introduce-legislation-to-ban-artificial-superintelligence-and-temporarily-pause-advanced-ai-development/

[...]
“Nearly every day, there is a frightening new story about how Big Tech companies are losing control of the technology they are developing, with potentially cataclysmic results,” Sanders said. “The leaders of the major AI companies publicly acknowledge that they do not fully understand the technology and that it is escaping their control. It is irresponsible for society to allow them to move forward and make these products even more advanced. That’s why I am introducing legislation to immediately pause the development of increasingly powerful AI and ban the creation of systems that humanity cannot fully control — at home and around the world. The future of humanity cannot be left in the hands of a handful of Big Tech oligarchs. The American people and people throughout the world must determine that future.”
“If we allow Artificial Superintelligence to be built, it could risk the security, freedom, and lives of Americans,” Casar said. “Despite its potential deadly consequences, cutting-edge AI technology is less regulated than the average food truck. That must change. In just four years, we have gone from the first version of ChatGPT to AI models so powerful they cannot be properly controlled. Congress should immediately ban AI systems too powerful to control.”
[...]
The Artificial Superintelligence Ban Act addresses these recent events and holds these frontier labs to their words by:
Banning AI superintelligence so no person or entity may develop or deploy Superintelligent AI systems that surpass human intelligence or have the capacity to overthrow human governments, or systems that have dangerous abilities like subverting shutdown commands.
Pausing advanced AI development until a new, federal AI regulatory body is up and running and has established clear rules and model review process to ensure safe and secure development and deployment of AI. 
Establishing a new cabinet-level federal agency to safeguard the public from the dangers of artificial intelligence, including by enforcing a prohibition on artificial superintelligence. This agency will be advised by an Artificial Intelligence Advisory Board comprised of experts on artificial intelligence to provide independent scientific and technical advice on matters related to artificial intelligence. The agency will: 
Monitor frontier AI systems at all stages of the lifecycle for dangerous capabilities. 
Supervise the removal of dangerous capabilities.
Supervise the destruction of artificial superintelligence.
Setting penalties for any person or entity that attempts to violate or circumvent the pauses and prohibitions laid out in this bill. Entities shall be subject to the corporate death penalty, and persons shall be subject to not more than 20 years in prison, which is similar to existing penalties related to unlawfully developing nuclear weapons.
Working to ban superintelligence around the world by setting the international policy of the United States to pursue international agreements, allied coordination, and policies such as export controls to prevent the development of artificial superintelligence anywhere in the world.
Read a summary here.

AI Development PauseAI GovernanceNewsPoliticsAI

Personal Blog

225

New Comment

31 comments, sorted by

top scoring

Click to highlight new comments since: Today at 9:23 PM

[-]DanB1mo2911

It is very ominous to me that both of these politicians are on the far left side of the political spectrum. This will tend to partisanize ideas about AI Safety, in a completely unnecessary way, since AI is very unpopular among the conservative base.

[-]Hastings1mo2314

This is a general argument that proves it's bad when leftists do good things, based on the claim-about-the-world that for any action they take, America as a collective will take action to not only cancel their action but go past zero in the opposite direction.

One might suspect that if this is the case, the problem isn't the leftists.

[-]MichaelDickens1mo51

IMO a plausible model is that once an issue already has determined partisan lines, marginal effort in one direction does not get canceled out. But if an issue isn't partisan yet, then efforts coming disproportionately from one side of the spectrum can have an outsized effect on making it partisan, which ends up being still better than nothing, but much worse than an effort that was politically balanced even if much less effective—e.g. if you can do some political activism that has a 10% chance of getting a Republican and a Democrat to co-sponsor a bill, vs. a 50% chance of getting to far-left Democrats, then the 10% chance might be better.

(I don't have a clear picture of what the right model is, but I find this model plausible enough that I'm worried about it.)

[-]DanB1mo1-1

One might suspect that if this is the case, the problem isn't the leftists.

Sure, the problem is the ambient Culture War. I wrote about how we need a Culture War ceasefire to address many goals, including AI safety, animal welfare and just fixing the national debt; any reasonable observer must view these goals as far more important than any of the petty priorities of the political conflict.

[-]Matrice Jacobine1mo38

How is this not just "we need a ceasefire on the political issues I don't care about to address the political issues I care about"? Animal rights and fiscal conservatism are extremely partisan (in opposing directions) and controversial!

[-]DanB1mo1-2

Animal rights and fiscal conservatism are extremely partisan

I don't see this. My criterion for whether an issue is partisan is whether politicians mention it in a speech attacking their rivals. I classify climate change, immigration, DEI, transgender rights, abortion, and gun rights as partisan issues on that basis. I don't see AI Safety or animal welfare as being in this category, and I separately recognize those issues as supremely important in the present moment.

There is strong nascent pro-animal-welfare sentiment on the right. Here are two tweets from arch-MAGA commentator Mike Cernovich:

Pigs are as emotionally aware as dogs are and a lot of 🐖 🐷 🐽 are raised in situations we would consider animal abuse and outright torture.

Eating pigs, by any objective measure, is equivalent to eating dogs. ... Pigs are actually more conscious than dogs are.

[-]Matrice Jacobine1mo20

https://time.com/article/2026/06/02/james-talarico-girlfriend-vegan-polls/

(Also, are you implying that politicians never mention the deficit in speeches attacking their rivals? What?)

[-]Raemon1mo7-3

Yeah I am worried about this but don't feel like I know enough to call the sign.

My uninformed take:

I think it probably would have been better to propose this in a few months when the conservatives have had more time to think "okay shit this AI stuff seems real, we need to do something" and come with some conservative flavored version of the thing, and then this might still cause partisanship but the right wing would have consolidated around some position that a) handled the problem at all, and b) is conservative flavored enough that the debate can be between two At Least Somewhat Helpful plans.

[-]Matrice Jacobine1mo11

From what I've seen, the Bannon-Allen right-populist argument has openly been "the Blues are eating us on this front, we need to do something" for a while. There was a touch of "Big Tech is woke" before Trump 2 but it's gone in favor of "the establishment GOP is betraying the MAGA base for money".

[-]FlorianH1mo5-3

If anything I'd believe in reverse causality: if it's so unpopular among the conservative base but only the far left is bringing it up, maybe a large range of non-far-left politicians are a bit in bed with the big companies. But even that I'm not sure.

I have no clue; my gut feeling simply is: even in a super partisan world your conservative base doesn't "oh Bernie is for halting let's not halt", neither consciously not subconsciously very much; ; and in that sense also politicians don't have a straightfwd reason to either.

[-]StanislavKrym1mo1-1

What if the conservatives don't actually believe in the goddamn threat? To them the AIs could be artificial idiots or tools usable for math, coding and hacking. Maybe we could ask them on their thoughts about robots doing things, then explain how the AGIs would be as capable of bossing the robot bodies as our brains boss our bodies?

[-]Raemon1mo111

That could be. But, I think the actual evidence is "republicans and democrats have been surprisingly comparable in terms of how much they're updating, with a couple outliers". (the graph of "senators speaking out about AGI was within the same rough ballpark for blue and red politicians)

[-]Joey Yudelson1mo193

Very excited to see this, and in particular heartened by how shocking Bernie seems to have found the HuggingFace incident.

Object level, here's a manifold market for what will happen to this act:

[-]Charlie Steiner1mo166

Probably great news. They got the spirit. Yes, I expect this to get bogged down and politicized. Oh well, probably inevitable.

This seems like a top 25% ban proposal relative to my expectations a few months ago, and I think is still in the bottom 35% for how politicized it makes the issue. I still expect the chance of passing to be low, but at least it's in the Overton window now.

[-]TsviBT1mo140

Is the bill itself good? (I mean this only in full context; so this may be a fairly low bar, like "at least states the goal" and "doesn't include any really bad stuff".)
Is this something that we (someone) should be supporting in an organized way? Like sending emails to reps, or phone banking to support the bill?

[-]gilch1mo112

I emailed my reps using the tool at https://controlai.org. Seemed like a good time.

[-]I.M.J. McInnis1mo20

Unknown as yet.
Yes, very much!! Call your elected representatives (repeatedly!) to tell them you're a single-issue voter on this topic.

[-]_rpd1mo10

the text of the draft bill is not yet publicly available

[-]Ben Pace1mo1110

Hurrah! That's the stuff! If more senators care and are brave like this, maybe me and everyone I love won't die.

[-]Gordon Seidoh Worley1mo85

I have low expectations that this specific law passes, but high expectations that this works to build consensus that we must do something. Even though there's reasonable risk of politicization of AI, it's hopefully counterbalanced by the broad unpopularity of data centers, including among the politically right (GOP seems quite worried about data centers costing them seats in the mid-terms), which means there's a way to sell this not as "liberals holding back progress" but "pushing the Overton window in a direction that will lead to something reasonable". My big hope is that this moves quickly enough that it influences the upcoming Trump/Xi meetings, and creates momentum around getting a treaty in place.

[-]MichaelDickens1mo53

I have low expectations that this specific law passes, but high expectations that this works to build consensus that we must do something.

On general principle, I am in favor of starting by proposing the bill you actually want, and then negotiating it down to the bill that you can get passed. This bill looks more like the thing I actually want than any other bill I've seen proposed.

[-]MichaelDickens1mo74

Great to see that the legislation includes serious penalties rather than the usual "your company will be fined an amount equivalent to 0.00001% of revenue."

[-]Mateusz Bagiński1mo43

Yes, please.

Pausing advanced AI development until a new, federal AI regulatory body is up and running and has established clear rules and model review process to ensure safe and secure development and deployment of AI. 

Q to anyone who read the thing: what does this imply for research on advanced/cutting-edge-ish, but narrow/specialized AI, e.g., in the ballpark of alphafold?

[-]Alexis Davis1mo2-1

This is the part that I am struggling with! I don't see how a pause will be when these advancements have been made without oversight, and have resulted in real scientific advancement, and of course unheard-of corporate gains for the companies leading the charge.

I feel like Sen. Sanders' proposal misses the mark on the economic and scientific implications of what is happening within these labs. Sure, yes, safe technological development 'should matter' the most. But that has not been the case. The summary sounds like a fear-mongering blanket ban... hoping that isn't the case (because regulation would be awesome), but it's not looking great.

[-]David Gross1mo31

The idea that you can reliably make something better by giving the U.S. federal executive branch more authority over it is an idea that has been reductus ad absurdum so thoroughly by now that I'm surprised it is still taken so seriously here.

I'm sympathetic to the idea that we need a pause in AI capabilities development, etc. but the theory that the path toward doing this effectively and well runs primarily through the White House doesn't seem to me to pass the laugh test.

[-]sanyer1mo20

How are they planning to define Superintelligence? I don't think it's enough to either define it as something that surpasses human intelligence (current systems already do in many domains), or something that the capacity to overthrow human governments (outcome-based AGI bans are not feasible). The third option, "systems that have dangerous abilities like subverting shutdown commands", seems more feasible, but there again, I don't think we have any good measures for this currently.

[-]_rpd1mo10

This appears to be the working definition ...

“Artificial Superintelligence” means:

An artificial intelligence system that exhibits or can easily be modified to exhibit capabilities that match or exceed human cognitive performance and capabilities across a broad range of domains or tasks.

Or AI systems that have sufficient capabilities to plan and execute the disempowerment of humanity, including by overthrowing or undermining the U.S. government.

https://www.sanders.senate.gov/wp-content/uploads/Ban-Artificial-Superintelligence-Act-Release-Summary.pdf

[-]Ustice1mo2-12

I‘d like to see them try to define ASI in a way that it can be recognized before it‘s too late to be enforceable.

I’d much rather they strongly encourage and support open source/weights/research/reports for AI. Kind of like a reverse Manhattan Project, where instead of secretly gathering top researchers, we get as many eyes and brains on it as possible working togetherish.

But we won’t do that; we’ll do something stupid and reactionary or nothing at all.

[-]aphyer1mo10

As someone who thinks that AI risk exists but is skeptical that government involvement will make things worse rather than better, I find the fact that the AI safety movement's standard-bearer in Washington seems to be Bernie Sanders rather discouraging.

[-]I.M.J. McInnis1mo01

Thrilling to see. Smashing open the Overton window like the Kool-Aid pan. Yes, I wish it foregrounded the internationalness of the ban, and yes, I wish there were a right-wing cosponsor, but let not perfect be the enemy of the good. This is a headline you see in a world that survives.

[-]Julian Bradshaw1mo-23

It's unclear to me if this is meant to ban superintelligence permanently? Ban until we're sure we have a working approach for aligning it, okay, but I don't like the idea of Butlerian Jihad entering the Overton Window... we do want the biological immortality/uploading/access to the stars/etc. that aligned ASI would help bring.

I mean, in a certain sense the ban as proposed is a little ill-defined since it might not cover speed-based superintelligence (running tons of AGIs which think way faster than humans), which we ought to get with just AGI after a bit, and that alone could plausibly bring the core benefits we want. I could imagine a sensible stance where we want to delay ASI a really long time and just rely on tons of AGI, though obviously there's still considerable disempowerment/x-risks even there, but it's not clear to me this proposal is based on such thinking.

Moderation Log