Headlines
On February 13, the Wall Road Journal reported one thing that hadn’t been public earlier than: the Pentagon used Anthropic’s Claude AI in the course of the January raid that captured Venezuelan Chief Nicolás Maduro.
It mentioned Claude’s deployment got here via Anthropic’s partnership with Palantir Applied sciences, whose platforms are broadly utilized by the Protection Division.
Reuters tried to independently confirm the report – they could not. Anthropic declined to touch upon particular operations. The Division of Protection declined to remark. Palantir mentioned nothing.
However the WSJ report revealed yet one more element.
Someday after the January raid, an Anthropic worker reached out to somebody at Palantir and requested a direct query: how was Claude really utilized in that operation?
The corporate that constructed the mannequin and signed the $200 million contract needed to ask another person what their very own software program did throughout a army assault on a capital metropolis.
This one element tells you every part about the place we really are with AI governance. It additionally tells you why “human within the loop” stopped being a security assure someplace between the contract signing and Caracas.
How large was the operation
Calling this a covert extraction misses what really occurred.
Delta Drive raided a number of targets throughout Caracas. Greater than 150 plane have been concerned. Air protection techniques have been suppressed earlier than the primary boots hit the bottom. Airstrikes hit army targets and air defenses, and digital warfare belongings have been moved into the area, per Reuters.
Cuba later confirmed 32 of its troopers and intelligence personnel have been killed and declared two days of nationwide mourning. Venezuela’s authorities cited a loss of life toll of roughly 100.
Two sources told Axios that Claude was used in the course of the lively operation itself, although Axios famous it couldn’t verify the exact position Claude performed.
What Claude would possibly even have accomplished
To grasp what may have been occurring, it’s essential to know one technical factor about how Claude works.
Anthropic’s API is stateless. Every name is impartial i.e. you ship textual content in, you get textual content again, and that interplay is over. There isn’t any persistent reminiscence or Claude working constantly within the background.
It is much less like a mind and extra like an especially quick advisor you possibly can name each thirty seconds: you describe the state of affairs, they offer you their greatest evaluation, you grasp up, you name once more with new data.
That is the API. However that claims nothing concerning the techniques Palantir constructed on high of it.
You’ll be able to engineer an agent loop that feeds real-time intelligence into Claude constantly. You’ll be able to construct workflows the place Claude’s outputs set off the following motion with minimal latency between suggestion and execution.
Testing These Situations Myself
To grasp what this really appears like in follow, I examined a few of these situations.
each 30 seconds. indefinitely.
The API is stateless. A complicated army system constructed on the API does not must be.
What which may appear like when deployed:
Intercepted communications in Spanish fed to Claude for fast translation and sample evaluation throughout a whole lot of messages concurrently. Satellite tv for pc imagery processed to establish automobile actions, troop positions, or infrastructure adjustments with updates each jiffy as new photos arrived.
Or real-time synthesis of intelligence from a number of sources – alerts intercepts, human intelligence stories, digital warfare knowledge – compressed into actionable briefings that might take analysts hours to provide manually.
skilled on situations. deployed in Caracas.
None of that requires Claude to “determine” something. It is all evaluation and synthesis.
However while you’re compressing a four-hour intelligence cycle into minutes, and that evaluation is feeding immediately into operational choices being made at that very same compressed timescale, the excellence between “evaluation” and “decision-making” begins to break down.
And since it is a categorised community, no person exterior that system is aware of what was really constructed.
So when somebody says “Claude cannot run an autonomous operation” – they’re in all probability proper concerning the API stage. Whether or not they’re proper concerning the deployment stage is a very totally different query. And one no person can at the moment reply.
Hole between autonomous and significant
Anthropic’s arduous restrict is autonomous weapons – techniques that determine to kill with no human signing off. That is an actual line.
However there’s an infinite quantity of territory between “autonomous weapons” and “significant human oversight.” Take into consideration what it means in follow for a commander in an lively operation. Claude is synthesizing intelligence throughout knowledge volumes no analyst may maintain of their head. It is compressing what was a four-hour briefing cycle into minutes.
this took 3 seconds.
It is surfacing patterns and suggestions sooner than any human crew may produce them.
Technically, a human approves every part earlier than any motion is taken. The human is within the course of. However the course of is now shifting so quick that it turns into inconceivable to guage what’s in it in quick paced situations like a army assault.When Claude generates an intelligence abstract, that abstract turns into the enter for the following determination. And since Claude can produce these summaries a lot sooner than people can course of them, the tempo of your entire operation quickens.
You’ll be able to’t decelerate to consider carefully a couple of suggestion when the state of affairs it describes is already three minutes previous. The data has moved on. The subsequent replace is already arriving. The loop retains getting sooner.
90 seconds to determine. that is what the loop appears like from inside.
The requirement for human approval is there however the potential to meaningfully consider what you are approving will not be.
And it will get structurally worse the higher the AI will get as a result of higher AI means sooner synthesis, shorter determination home windows, much less time to assume earlier than performing.
Pentagon and Claude’s arguments
The Pentagon wants access to AI models for any use case that complies with U.S. legislation. Their place is basically: utilization coverage is our downside, not yours.
However Anthropic desires to keep up particular prohibitions – no totally autonomous weapons and prohibiting mass home surveillance of Individuals.
After the WSJ broke the story, a senior administration official instructed Axios their partnership/settlement was below overview and that is the explanation Pentagon said:
“Any firm that might jeopardize the operational success of our warfighters within the area is one we have to reevaluate.”
However sarcastically, Anthropic is at the moment the one industrial AI mannequin permitted for sure categorised DoD networks. Though, OpenAI, Google, and xAI are all actively in discussions to get onto these techniques with fewer restrictions.
The true battle past arguments
In hindsight, Anthropic and the Pentagon could be lacking your entire level and pondering coverage languages would possibly resolve this subject.
Contracts can mandate human approval at each step. However, that doesn’t imply the human has sufficient time, context, or cognitive bandwidth to really consider what they’re approving. That hole between a human technically within the loop and a human really in a position to assume clearly about what’s in it’s the place the actual danger lives.
Rogue AI and autonomous weapons are in all probability the later set of arguments.
In the present day’s debate needs to be – would you name it “supervised” while you put a system that processes data orders of magnitude sooner than people right into a human command chain?
Remaining ideas
In Caracas, in January, with 150 plane and real-time feeds and choices being made at operational pace and we do not know the reply to that.
And neither does Anthropic.
However quickly, with fewer restrictions in place and extra fashions on these categorised networks, we’re all going to search out out.
All claims on this piece are sourced to public reporting and documented specs. We have now no private details about this operation. Sources: WSJ (Feb 13), Axios (Feb 13, Feb 15), Reuters (Jan 3, Feb 13). Casualty figures from Cuba’s official authorities assertion and Venezuela’s protection ministry. API structure from platform.claude.com/docs. Contract particulars from Anthropic’s August 2025 press launch. “Visibility into utilization” quote from Axios (Feb 13).
