The Secretary of War didn’t really mean it, contends US government lawyer as Anthropic gets its first day in court over Trump 2.0’s risk to the nation designation. So what did he mean?

So how did Anthropic’s first day in court facing off against the US Department of War (DoW) go? Well, we won’t find out for a few days yet, but presiding District Judge Rita Lin of the California Northern District Court was very clear about what she sees as at being at stake here - and made clear that she doesn’t see her role as being “to decide who is right” in the “fascinating public policy debate” around ethical AI use.

But there is still a lot riding on the outcome of her decision on whether to grant Anthropic an injunction to block the blanket ban put in place by Trump 2.0  on use of its tech, a ban which the supplier claims would cost it “multiple billions of dollars” and is an over-reach of government authority. Lin noted:

One of the amicus briefs [filed by third parties in support of Anthropic] used the term ‘attempted corporate murder’. I don’t know if it’s murder, but it looks like an attempt to cripple Anthropic. And specifically my concern is whether Anthropic is being punished for criticizing the government’s contracting position in the press.

She added:

After Anthropic went public with this contracting dispute, defendants seemed to have a pretty big reaction to that...What is troubling to me about these reactions is that they don’t really seem to be tailored to the stated national security concern.

What did he mean?

In court, Anthropic lead counsel was Michael Mongan from law firm WilmerHale, while Deputy Assistant Attorney General Eric Hamilton appeared for the defence on behalf of the government. In what was a neat and concise hearing, Lin ran through a series of questions she wanted answers to, beginning with a tweet sent out by Secretary of War Pete Hegseth on 27 February that announced:

Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic...This decision is final.

Now, several weeks later, she wanted to know, does the DoW contend that this is a legitimate legal order? Hamilton performed a retreat on that one, admitting that Hegseth’s comment had no legal authority and is not an agency action. It was just a social media post announcing that an action was being taken and that’s how everyone ought to have understood it, he insisted. 

So when the Secretary of War said, without being prompted, that this was his final decision and was what is going to happen, this was not in fact a “correct statement of what the DoW is about to do”, clarified Lin. No, it isn’t, admitted Hamilton.

That obviously begs the immediate question, which Lin was quick to ask, of how Anthropic or anyone else is supposed to know that, even though Hesgeth said it in public, he didn’t really mean it?

The DoW line is that the comment was followed up by a letter several days later making clear what the official position was, while further information was provided in court filings, although, of course, timing-wise that only happened after Anthropic began legal action.

So, does this mean that the US Government now concedes that Hegseth exceeded his authority with that social  post, asked Lin:

Why did Secretary Hegseth say this if it has no legal effect and he didn't intend it to happen?

For the defence, Hamilton said he didn’t know.

Damage

President Donald Trump had weighed into the row between Anthropic and the DoW with a blanket ban on the US Government doing work with any third party using Anthropic tech, a sweeping statement that left a lot of questions in its wake about how far this applied up and down the supply chain. Lin wanted her own clarification here:

Suppose i'm a contractor who supplies toilet paper to DoW. I’m not going to be terminated for using Anthropic?

The DoW still has concerns about sub-contractors using Anthropic tech for DoW work, but not non-DoW work, at this point, replied Hamilton.

That being so, what does that mean for Anthropic’s claim that the US Government’s actions risk costing it billions as customers and partners hesitate to work with it? For the plaintiffs, Mongan said the row had caused “immediate harm to economic and reputational interests”. 

While conceding that the DoW had indeed sent a letter to Anthropic to set out its actions in more detail, he pointed out that this did not arrive until 20 days after Hegseth’s now deemed invalid public pronouncement. He also argued that one reason why Anthropic needs "injunctive relief" and "authoritative clarity" from the court is that Hegseth has not publicly withdrawn his comment.

But when asked by Judge Lin what there is to stop the DoW from changing its view on this, Hamilton replied that DoW isn’t going “to change its interpretation of the social media post”.

Too much?

Attention also turned to whether the DoW’s actions have been excessively punitive. Lin noted that that there’s a difference between the DoW stopping working with Anthropic and deeming it overnight to be a supply chain risk to the security of the nation, using language involving sabotage and adversaries usually reserved for terrorist organizations and hostile foreign powers.

This is all hypothetical, admitted Hamilton, conceding that the DoW position here is based on concerns about what might happen at some future date rather than anything that actually has occurred. But, for example, what might happen when the tech needs upgrades, he wanted to know:  

Sabotage could be introduced through the updates.

So, the implication here is that the DoW has acted as it has on a hunch? Or, as Lin returned to one of her most basic questions, is this really all about the DoW just being annoyed because Anthropic has been “stubborn” in not acceding to government demands to ditch its ethical use red lines and “asks annoying questions”?

Acting stubbornly is not enough to justify such a risk designation, admitted Hamilton, but he insisted Anthropic is doing more than that, and that as a result in raising concerns about how DoW uses its technology. It all comes down to trust, he argued, and Anthropic has destroyed that trust and made it impossible to be regarded as “reliable and trustworthy partners” by raising questions around lawful uses of its technology:

If Anthropic is pushing back now, what's going to happen in the future, when our warfighters need it most? That's unacceptable risk.

The imposition of the sweeping supply chain risk designation was done for ease and efficiency, he went on, providing “one tool to address a risk” that could be used across agencies, rather than having to go through individual contracts and designate risk via what Hamilton called “a patchwork approach” that is “not a preferred option in the context of a national security risk”.

To which claim, Mongan hit back:

This is a supply chain risk designation in search of a rationale.

My take

Whether Lin sees DoW’s behavior as a ‘sledgehammer to crack a nut’ approach or a valid strategy seems likely to be at the kernel of her ruling later in the week. For now, we can only wait to see what happens next.

More to come.