Anthropic Takes a Stand

Politics & Current Affairs
1

The Washington Post reports that Anthropic rejects Pentagon terms for lethal use of its chatbot Claude. The dispute has arisen over the terms and conditions for the Pentagon’s use of Claude.ai for military and surveillance purposes.

Defense Secretary Pete Hegseth said the Pentagon must be able to use the technology for the full range of warfighting — a broad remit that left too many questions for Anthropic to be comfortable with.

On Thursday, Anthropic CEO Dario Amodei said in a lengthy statement that the firm was holding firm to its red lines — and hoped the Pentagon would reconsider.

“In a narrow set of cases, we believe AI can undermine, rather than defend, democratic values,” Amodei wrote. “Some uses are also simply outside the bounds of what today’s technology can safely and reliably do. Two such use cases have never been included in our contracts with the Department of War, and we believe they should not be included now,” he said, citing specifically autonomous weapons use and mass surveillance.

“We cannot in good conscience accede to their request,” Amodei wrote.

This comes on top of earlier reports that:

Anthropic’s artificial-intelligence model Claude was used in the U.S. military’s operation to capture former Venezuelan President Nicolas Maduro, the ​Wall Street Journal reported on Friday, citing ‌people familiar with the matter.

Claude’s deployment came via Anthropic’s partnership with data firm Palantir Technologies , whose platforms are widely used by the Defense Department and federal law enforcement, the ‌report ​added.

It’s not hard to imagine where this could end up, with ultra-smart AI technology, possibly including robotic agents in future, allied with Palantir’s deeply-embedded surveillance technology. Terminator springs immediately to mind.

Secretary Hegseth, true to form, is fuming and threatening to altogether cut Anthropic from government contracts among other things.

I think Amodei’s stance here deserves applause, as he’s up against an authoritarian and unprincipled administration who have already engaged in many questionable acts and demonstrated disregard for law and principle. As CEO of one of the leading AI firms, I think it’s encouraging that he’s taking a principled stand, at some political and business risk. Let’s hope his peers at other AI companies demonstrate similar commitment in future.

2 Likes
2

Agreed, and credit where it’s due — Amodei is putting real skin in the game here, not just issuing vague platitudes about “responsible AI.” That matters.

But I think its worth flagging something that tends to get lost in these stories. The fact that this comes down to one CEO’s conscience is itself the problem. What we’re watching is a situation where the only thing standing between a surveillance-military apparatus and basically unconstrained use of frontier AI is… a corporate policy decision. That’s not a stable equilibrium. Amodei could change his mind, or get replaced, or Anthropic could get acquired. The “red line” here is drawn in sand, not carved in stone.

The deeper issue is that we don’t have any institutional framework — legal, political, or otherwise — that can reliably constrain how these tools get deployed once they exist. And thats not an accident; its a feature of how the current administration relates to power. The whole posture is: capability confers permission. If we can do it, we should do it, and anyone who disagrees is an obstacle.

Which is exactly the conflation that needs to be resisted — the slide from “technically possible” to “operationally useful” to “morally permissable.” Those are three very different judgments, and collapsing them is how you end up in dark places fast.

So yeah, applause for Amodei. But the real question is whether we can build something more durable than one man’s good conscience.

1 Like
3

Going on history, I think probably not - the pace of development of the internet completely over-ran any legislative restraint for decades. And AI is happening even faster. Which in a way, is all the more reason to commend Amodei. Him standing up is about the only thing that could be done at this juncture.

4

The drive to keep pace with foreign powers provides constant pressure to misuse new tech. And freedom of companies to operate in different countries means that no one country can regulate them totally. The solution to both problems is a democratic world government, perhaps starting with the UN.

5

In other news… Man who makes deal with Leopard not to eat his face has face eaten.

2 Likes
6

Oh yeah, Donald Trump and Pete Hegseth just :heart_suit: the U.N.

Not.

7

I’m not sure who’s right on this issue but I strongly suspect that if the President of the moment happened to be from the other party and yet his policies on this matter were the exact same in every way then you’d be going on about how Anthropic are far right Russia-sympathizing insurrectionists undermining our democracy.

Anthropic could be right. Or the Pentagon could be right. But either way, we should not be judging this issue based on how much we happen to like or dislike the current President. If the other guys were in office, there’s a good chance the Pentagon would be doing the exact same stuff. Presidents come and go. We don’t suddenly get a completely different military just because you lost one election.

8

Predictably, I guess, the Administration has issued a rebuke to Anthropic, along with a DD ban from dealing with them, declaring them a Supply Chain Risk, which is a drastic step.

I quote:

“The Leftwing nut jobs at Anthropic have made a DISASTROUS MISTAKE trying to STRONG-ARM the Department of War, and force them to obey their Terms of Service instead of our Constitution,” Trump wrote in a post on his social media site Truth Social, using the administration’s preferred name for the Defense Department. “Their selfishness is putting AMERICAN LIVES at risk, our Troops in danger, and our National Security in JEOPARDY.”

The company responded:

“Designating Anthropic as a supply chain risk would be an unprecedented action — one historically reserved for US adversaries, never before publicly applied to an American company.” The unsigned blog post added: “We are deeply saddened by these developments.”

The Washington Post added:

The supply-chain-risk designation issued by Hegseth late Friday was an extraordinary escalation, ranking a leading American AI company alongside the likes of Chinese and Russian firms seen as a danger to the United States.

Meanwhile, Musk chimed in:

Musk’s xAI could also benefit from Anthropic’s dispute with the Pentagon. Defense officials have said xAI has already agreed to the Pentagon’s terms for working on classified systems. The entrepreneur jumped on Michael’s social media thread Friday, saying “Anthropic hates Western Civilization.”

One shudders to think what he could do with AI, considering the amount of damage he was able to wreak with a comparatively unsophisticated chainsaw.

Full article can be read here (gift link).

9

But are the Republicans going to lose another election. Or will it be stolen like the 2020 election.

10

Different topic. But I imagine it will be a subject for discussion.

11

Dario Amodei answers questions

12

It’s important, here, that Anthropic has done extensive work with the Department of Defense War but that they said the use of their tools for mass surveillance and autonomous weapons systems needed more discussion including with Congress.

For that impudence they were vilified and black listed by Trump and Hegseth.

‘Defying the Will of the Emperor’.

13

I disagree that this is at all ‘important here’. We’re talking about the development of a new, untested, but potentially powerful tool in warfare.

This is akin to the development of nuclear bombs, or biological weapons.

The fact that some imbecile in the White House threw a hissy fit over it is a sideshow.

The issue is that our exposure to new lethal military capabilities is, once more, left in the hands of a couple of CEOs of entirely private corporations and conducted behind closed doors with virtually zero public accountability.

This is not something that’s arrived with the latest dotard of a president. It’s something that’s been with us since MK Ultra.

14

It seems Anthropic is now officially a supply chain risk.

Is it over for Anthropic? If someone who understands what all of this means could lay it out in simple terms for me. Is it a really big deal? can Anthropic fight legally to avoid the catastrophe?
Was this all Sam’s plan? In which case the guy’s a genius!

15

Being categorised a ‘supply chain risk’ is apparently a drastic step and has never been applied to a US company before. It’s usually reserved for foreign thread actors It is typical of Trump’s petulant and vindictive attitude - ‘if you don’t do what I want, then you’ll suffer for it’. I don’t think it will drive Anthropic out of business, media reports are that most outside MAGA are on their side. I’ve just signed up for their paid sub having been a casual user for two years, and I can attest that Claude.ai is a superb product for my purposes.

Don’t loose sight of the fact that Anthropic’s only objections were to use of their product for mass surveillance and autonomous weapons systems. They say, quite rightly, that such uses should be reviewed by Congress. But Trump and Hegseth view congress with contempt.

16

And so it goes.

AI is poised to be used as a weapon and the news is still about Trump’s circus.

I swear someone could probably commit genocide and the world would still be talking about nothing but Trump… Oh no wait… That actually happened.

17

"The Trump administration’s push to punish the artificial intelligence startup Anthropic is undercutting the United States’ plan to boost American technology and dominate its spread worldwide, according to industry advocates and a former top Trump adviser on AI.

Fears of further escalation in the fight between the Pentagon and Anthropic over limits on military use of AI have already sparked major uncertainty for tech companies working with the government, roughly a dozen industry representatives, investors and former government officials told POLITICO.

They said the sector would suffer even greater damage if the administration designated Anthropic as a risk to the military supply chain — an action the Defense Department formally took Thursday. Anthropic has said it would fight any such designation in court.

Concerns over the impact of such an unprecedented action against a U.S. company reached a new pitch this week, as four top tech lobbying groups urged President Donald Trump to reconsider his administration’s supply-chain threat.

But Trump showed no signs of easing up, telling POLITICO in an interview earlier Thursday that Anthropic would face setbacks as a result of its dispute with the Pentagon.

“Anthropic is in trouble because I fired them like dogs, because they shouldn’t have done that,”

Speaks volumes.

From Politico

18

A well functioning AI is great for a police state. I think a good example is how the Chinese are using it. Earlier police states have simply collapsed under the sheer reality that you simply cannot have every being checked by someone in the security system even if vast networks of informers are established.

Yet the simple answer then is that people simply change their behaviour and speak openly only to the most trusted people in person when that surveillance is everywhere.

Good luck to the company in court. They still aren’t Kangaroo courts yet.

1 Like
19

Anthropic are also getting a great deal of support and sympathy from industry peers and the (non-MAGA) public. I think it’s safe to say that Pete Hegseth’s preposterous egotism and militarism are thoroughly repellent to many in Silicon Valley.

20

Anthropic is suing the Trump administration over the ‘supply chain-risk’ label according to this article.

Anthropic filed two federal lawsuits on Monday against the Trump administration alleging that Pentagon officials illegally retaliated against the company for its position on artificial intelligence safety.

The lawsuits, filed in the U.S. District Court for the Northern District of California and the federal appeals court in Washington, D.C., allege the Trump administration violated the company’s First Amendment rights and exceeded the scope of supply chain risk law by using the label against Anthropic. The suit is asking a federal judge to block Pentagon officials from enforcing the blacklist designation.

1 Like