Anthropic is blacklisted by the Pentagon and being used by the NSA at the same time

[midian182]

Facepalm: It appears that the US government isn't completely against American agencies using technology from Anthropic, a company it designated a "supply chain risk" earlier this year. According to a new report, the NSA is using the AI firm's Mythos model "for offensive cyber operations."

The Financial Times reports that Anthropic has installed half a dozen engineers inside the NSA as forward-deployed staff. Their job is said to involve guiding the agency's use of Claude Mythos and customizing the model for specific applications.

It remains unclear whether the Anthropic employees are helping with live hacking operations, or if Mythos is being used in active campaigns. But one person close to the arrangement told the FT that the model would be useful for infiltrating networks in countries such as China and Iran. The logic is that US adversaries are likely to be using cyber-focused AI systems of their own, making an early advantage important.

The report fits the NSA's remit, which includes collecting signals intelligence through wiretaps, undersea cables, corporate partnerships, and other clandestine methods, as well as offensive cyberattacks against foreign adversaries. The agency declined to confirm or deny the report, while Anthropic did not comment.

In February, the Pentagon threatened to cut Anthropic off unless it dropped restrictions on using Claude for mass domestic surveillance and fully autonomous weapons. The dispute reportedly began while Anthropic and the Department of Defense were negotiating a $200 million contract, before the administration demanded broader access to Claude for "all lawful purposes."

Anthropic refused, leading the government to label it a supply chain risk, putting it in the same broad penalty box as firms such as Huawei and ZTE. The company later said it would sue, arguing that the designation was legally unsound.

A federal judge in California temporarily blocked one version of the designation in March, but a Washington appeals court allowed the Pentagon blacklist to remain in place in April. By May, the Pentagon had signed major AI deals with OpenAI, Google, Microsoft, and Nvidia while cutting Anthropic out.

The NSA sits under the Department of Defense, the same department arguing in court that Anthropic's technology poses a national security risk. But reports suggest the NSA was already using Mythos despite the blacklist, implying parts of the DoD may consider the model too useful to ignore.

Anthropic has framed Mythos as a defensive cybersecurity tool. The company launched Project Glasswing in April with partners including AWS, Google, Microsoft, Nvidia, and CrowdStrike, saying Mythos had identified thousands of zero-day vulnerabilities.

This week, Anthropic said partners had found more than 10,000 high- or critical-severity flaws, and access was expanding to around 150 more organizations in over 15 countries.

Permalink to story:

Anthropic is blacklisted by the Pentagon and being used by the NSA at the same time

 

[Theinsanegamer]

I mean that sounds about on par for this administration

 

[Win7User]

NSA were caught red handed spying on Americans some time ago. Nothing has changed, these fools do what ever the hell they want to do.

 

[Fimbulvetr]

So Anthropic actually has some ethics, while OpenAI, Google, Microsoft, and Nvidia couldn't care less and will happily sell us all down the river...

 

[JaredTheDragon]

NSA were caught red handed spying on Americans some time ago. Nothing has changed, these fools do what ever the hell they want to do.

That is literally their only function, along with their pals in Langley.

 

[Endymio]

So Anthropic actually has some ethics, while OpenAI, Google, Microsoft, and Nvidia couldn't care less and will happily sell us all down the river...

Refusing the US military while doing business with CCP China is the exact opposite of "ethics".

Despite the clickbait headline, there's nothing contradictory about this situation. The DoD can't depend on Anthropic to plan military operations, but cyber ops are an entirely different task.

 

[Anton Longshot]

I have a dream.
My wish is that one day AI will effectively replace the bad and the stupid because those are the reason we don't and cannot have the really nice things.
I'm hoping politicians will be the first to go as those display dismal performance, without exceptions.

 

[passwordistaco]

Refusing the US military while doing business with CCP China is the exact opposite of "ethics".

Despite the clickbait headline, there's nothing contradictory about this situation. The DoD can't depend on Anthropic to plan military operations, but cyber ops are an entirely different task.

Anthropic apparently does not accept accounts from China or companies that are over 50% Chinese owned.

https://www.anthropic.com/news/updating-restrictions-of-sales-to-unsupported-regions

 

[Endymio]

Anthropic apparently does not accept accounts from China or companies that are over 50% Chinese owned.

Sure: they made that change last year, one month after signing their deal with the US DoD. Now that this deal no longer exists, will they revert to their old habits?

 

[VitalyT]

Right hand doesn't know what the left one is doing, sums up the entire administration.

 

[Endymio]

Right hand doesn't know what the left one is doing, sums up the entire administration.

I realize you don't care that the facts contradict your anti-Trump diatribes, but Trump's original directive labeled Anthropic a "supply chain risk", which merely bars its use from critical military or infrastructure systems, not offensive cyber ops. Not that it's even relevant, because after that Trump signed another directive allowing federal agencies to vet these AI models for national security concerns ... and that includes Anthropic:

Trump signs executive order that allows voluntary federal vetting of top AI models for national security risks

It was not immediately clear to what extent the order signed Tuesday differed from the one he declined to sign on May 21.

www.pbs.org

 

[Beerfloat]

Refusing the US military while doing business with CCP China is the exact opposite of "ethics".

There is no 'refusing the US military' here. Anthropic were perfectly happy to do business. The military also isn't the problem.

When the administration chooses to put our soldiers in harm's way in its illegal wars, unlawful domestic policing deployments, orders them to perform extrajudicial murder, uses them to threaten invasion of what used to be considered our allies, or engages them in election interference in said former alies, while at the same time performing unprecedented political meddling in the careers of qualified officers, the military loses. It becomes a victim of the administration's corruption just like everyone else. The contrast with the CCP, which should be obvious, becomes murkier.

 

[Endymio]

There is no 'refusing the US military' here. Anthropic were perfectly happy to do business.

Anthropic refused to allow its products to be used for military purposes. That makes them useless to the military. An AI model doing mission planning has to recalculate asset deployment in real time: what targets to strike in which priority based on rapidly-changing intel. Anthropic refused to allow it. Their perogative ... but if the tools are useless for the purposes given, then you can't expect the government to pay for them.

When the administration chooses to put our soldiers in harm's way in its illegal wars, unlawful domestic policing deployments, orders them to perform extrajudicial murder

These drama queen histrionics aren't fooling anyone. I doubt even you yourself believe this nonsense, it's merely your way of virtue-signaling hatred of Trump.

 

[Beerfloat]

Anthropic refused to allow its products to be used for military purposes. That makes them useless to the military. An AI model doing mission planning has to recalculate asset deployment in real time: what targets to strike in which priority based on rapidly-changing intel. Anthropic refused to allow it. They’re perogative ... but if the tools are useless for the purposes given, then you can't expect the government to pay for them.

Obviously the vast majority of military purposes do not involve autonomous killing machines or domestic surveillance. Sadly this administration has decided that Anthropic must be punished and excluded from all of them, regardless of whether they might actually have the best offering for any given task. A decision that appears to be driven by spite rather than any clarity of purpose.

These drama queen histrionics aren't fooling anyone. I doubt even you yourself believe this nonsense, it's merely your way of virtue-signaling hatred of Trump.

All indisputable facts. So of course you would not argue against them but rather misdirect. Funny you should mention Trump. You seem to have adopted one of his methods.

 

[Endymio]

Obviously the vast majority of military purposes do not involve autonomous killing machines or domestic surveillance.

Nonsense. By Anthropic's definition, we've had "autonomous killing machines" since WW2. You're arguing a truly absurd position: if an AI model can't direct military assets, then it's useless to the military. What else is it going to do -- plan healthier lunch menus for the troops?

I can imagine people like you during WW2, literally cheering on Anthropic as it tells the US military it can't use its products to help defeat the Nazis.

Sadly this administration has decided that Anthropic must be punished and excluded from all of them,

You're spreading disinformation. As the link I posted shows, Anthropic's models are currently being evaluated for several government -- including DoD -- uses.

 

[Beerfloat]

I can imagine people like you during WW2, literally cheering on Anthropic as it tells the US military it can't use its products to help defeat the Nazis.

It didn't require much prompting to make you go there, did it? So, other than easily triggered, where would you see yourself in this scenario? Maybe you'd be the guy busy scolding everyone who doesn't want to be the first to deploy some biological, chemical or nuclear weapons. The end seems to justify any means.

You're spreading disinformation. As the link I posted shows, Anthropic's models are currently being evaluated for several government -- including DoD -- uses.

Nonsense. Your story says nothing about revoking Anthropic's designation as a supply chain risk.
It's about how this mercurial administration is now considering some token AI regulation, which will be voluntary, brief, and understaffed because it has gutted most departments of anything resembling expertise.

 

[Endymio]

Maybe you'd be the guy busy scolding everyone who doesn't want to be the first to deploy some biological, chemical or nuclear weapons. The end seems to justify any means.

If "the ends" were defeating the Nazis earlier and preventing much of the Holocaust, as well as defeating the Japanese and their even worse atrocities in China, Indonesia, and elsewhere in Asia -- then most certainly I'd be calling for the use of any and all weapons in our arsenal. Maybe you should examine some photos from Dauchau or the Rape of Nanking before bleating any synthetic moral pretensions about preventing that sort of thing.

Nonsense. Your story says nothing about revoking Anthropic's designation as a supply chain risk.

You failed to read the link:

"The [executive] order establishes a framework for the federal government to vet the national security risks of advanced AI systems... That directive was characterized as a voluntary collaboration with participating U.S.-based tech companies, including Anthropic, OpenAI and Google, ...

The White House said in a social media post Tuesday that the executive order "creates a process for [these] labs to voluntarily share cutting-edge cyber models in order to secure critical infrastructure and strengthen the government's own cyber defenses...."

The designation still exists, but individual agencies can vet a model against their own specific needs and determine if a national security risk exists in that particular case.

 

[Northy]

I hope one day I can have my own AI murder bot too, will call it "Peace Lovin' Gary".

 

[Beerfloat]

If "the ends" were defeating the Nazis earlier and preventing much of the Holocaust, as well as defeating the Japanese and their even worse atrocities in China, Indonesia, and elsewhere in Asia -- then most certainly I'd be calling for the use of any and all weapons in our arsenal. Maybe you should examine some photos from Dauchau or the Rape of Nanking before bleating any synthetic moral pretensions about preventing that sort of thing.

Nothing synthetic about it - the US is a signatory to treaties banning the use of biological and chemical weapons, as well as nuclear non-proliferation and disarmament, because collectively as societies we have developed the ability to learn from mistakes and aim to make improvements.

Conversely, and unsurprisingly, here we find you once again advocating for war crimes, showing that same lack of imagination that has you fiercely condemning Anthropic for not wishing to be involved with autonomous killing machines. It seems that even if you had the benefit of similar hindsight for this particular class of weapons, you'd be the guy who would still choose to press full steam ahead.

 

[Endymio]

the US is a signatory to treaties banning the use of biological and chemical weapons, as well as nuclear non-proliferation and disarmament ... Conversely, and unsurprisingly, here we find you once again advocating for war crimes,

Now you're just being absurd. There were no "nuclear non-proliferation or disarmament treaties" In 1945 WW2, nor was our use of -- my exact phrase -- "any and all weapons in our arsenal" a war crime.

What *was* an war crime under international law, however, was the Biden Administration's decision to send cluster bombs to Ukraine, weapons banned under the 2008 Convention on Cluster Munitions, and weapons the US has widely criticized other nations for using.. Where was your synthetic outrage then? Why not admit you don't care about the actual details here, you simply automatically disagree with Trump on any and all issues?

...you fiercely condemning Anthropic for not wishing to be involved with autonomous killing machines

You're confused. If Anthropic truly didn't "wish to be involved" with military use of its technology, it wouldn't now be suing the federal government to force the DoD to use their products for military purposes. They want to have their cake and eat it too.

 

[calculatorsking1]

It may seem contradictory at first, but government agencies often have different policies, requirements, and risk assessments. A company can face restrictions or scrutiny from one part of the government while still providing services to another. The important question is what specific policies, contracts, or security concerns are driving those decisions rather than assuming all agencies take the same approach.

 

[Beerfloat]

What *was* an war crime under international law, however, was the Biden Administration's decision to send cluster bombs to Ukraine. Where was your synthetic outrage then, eh?

Your need to keep signalling your hatred of Biden is starting to border on the pathological. So here we see you clutching at yet another straw man of your own creation. I reckon you're actually aware that the US is not a signatory to any treaty banning cluster munitions, and neither are Russia or the Ukraine. And indeed Russia has been using them indiscriminately throughout their aggression against Ukraine. So what exactly would you like Ukraine to do here? Disarm unilaterally? Is there any limit to your ability to be on the wrong side of history on almost every single topic?

You're confused. If Anthropic truly didn't "wish to be involved" with military use of its technology, it wouldn't now be suing the federal government to force the DoD to use their products for military purposes. They want to have their cake and eat it too.

Anthropic is suing against the retaliatory designation as a supply-chain risk. I'm not sure why you would consider this at all confusing. They're looking to bid on government contracts just like their competitors. Why shouldn't they?

 

[Endymio]

Your need to keep signalling your hatred of Biden is starting to border on the pathological

Are you intentionally trying to embarrass yourself? Long before I brought up Biden in this thread, you were making these pathologically disturbed remarks about the Trump Administration. It's clear who has the problem here.

I reckon you're actually aware that the US is not a signatory to any treaty banning cluster munitions

Oops! 120+ nations -- the majority of the world -- has banned these weapons: they are de facto illegal under International Law, and in fact the US has in the past roundly condemned their use.

Even worse is that neither the US nor ANY nation in the world is "signatory to a treaty" banning the use of AI in military weapons systems. Yet you've repeated claimed this to be illegal and immoral, and applauded Anthropic's "ethics" for their stance here. Contradict yourself much?

.... And indeed Russia has been using them indiscriminately throughout their aggression against Ukraine. So what exactly would you like Ukraine to do here? Disarm unilaterally?

Ah, sounds like you're espousing "the ends justify the means" philosophy you decried earlier. It's OK for Ukraine use these internationally banned weapons ... but using a nuke to end the Holocaust and the Japanese genocide in Manchuria is evil? Is there no limit to your innate hypocrisy?

 

[Endymio]

Anthropic is suing against the retaliatory designation as a supply-chain risk. I'm not sure why you would consider this at all confusing.

Nothing confusing about it. Anthropic is demanding the DoD buy their products, while simultaneously banning the DoD from being able to use them effectively.

They're looking to bid on government contracts just like their competitors. Why shouldn't they?

They're freely able to bid on any government contracts that don't involve the DoD or critical national security systems. And -- as my earlier link demonstrates -- even that is subject to review, as each individual federal agency can now vet Anthropic against their particular needs.

 

[Beerfloat]

Are you intentionally trying to embarrass yourself? Long before I brought up Biden in this thread, you were making these pathologically disturbed remarks about the Trump Administration. It's clear who has the problem here.

The operative words being 'in this thread', because you keep bringing him up, for reasons.
While happily running interference for the many transgressions of the actual, current administration.

Ah, sounds like you're espousing "the ends justify the means" philosophy you decried earlier. It's OK for Ukraine use these internationally banned weapons ... but using a nuke to end the Holocaust and the Japanese genocide in Manchuria is evil? Is there no limit to your innate hypocrisy?

You could've just said 'yes'. Yes, you think that Ukraine should not respond to Russian's use of cluster munitions to kill and maim its people.