The clash between Anthropic and the Pentagon is still unfolding — with the war against Iran, Sam Altman, and enormous sums of money all entangled in one of the most significant philosophical questions the world has faced in recent memory.
-
First, a summary of events:
Anthropic, the company behind the popular AI model Claude, lost a $200 million Pentagon contract after refusing to remove restrictions it had embedded in its model.
The Pentagon transferred the contract to OpenAI and designated Anthropic a national security risk — effectively barring it from signing commercial agreements with a long list of companies connected to the U.S. military. In response, Anthropic filed a lawsuit against the Pentagon, alleging that it had overstepped its authority and was acting out of retaliation.
-
A closer analysis of the case reveals some particularly interesting dimensions.
First, Anthropic's defiance of the government is a commercial move, designed to signal trustworthiness to private and business customers around the world — which is precisely what happened, as the Claude app surged up the download charts in the wake of the incident.
The company may have taken a page from Apple's playbook: Apple previously refused to give security agencies access to intelligence-valuable data encrypted on a locked iPhone.
The Pentagon's punitive response was designed to turn that kind of move into a financial loss, so that companies considering government defiance as a marketing tool will think twice before doing so.
It's reasonable to assume that Sam Altman of OpenAI took the Pentagon deal because he had little real choice.
Although many users around the world rushed to delete the ChatGPT app following the agreement, it's worth remembering that the company is in a precarious financial position — facing intensifying competition in the consumer market and a limited footprint in the enterprise and government sectors. Signing the Pentagon deal is a lifeline that could also open the door to further government contracts of significant scale.
-
But the most fascinating aspect of this story is the peculiar role reversal at its heart.
A private company, controlled by a small group of businesspeople, is attempting to impose regulatory constraints on the United States government — no less.
From a marketing standpoint, many users undoubtedly see the move as protecting the ordinary user from the power of the state. But in reality, the situation is precisely the opposite.
Governments have procedures, oversight mechanisms, and public legitimacy.
Businesspeople seeking to maximize profit through moves like this have none of those things — and government regulation is the only guardrail preventing them from causing far greater harm.
One need only look at the destructive, deliberate, and sustained impact that various social media platforms have had on the mental health and social fabric of people across the globe to understand just how dangerous business interests can be when left without adequate regulation.
It's also worth remembering that Anthropic's models played a significant role in the most recent round of fighting against Iran, serving as a meaningful force multiplier that identified missile launchers and military targets — and in doing so, saved many lives.
Furthermore, it's important to keep in mind that terrorists, criminal organizations, and governments such as those of China and Russia are using artificial intelligence without any restrictions whatsoever to advance highly dangerous objectives against the West. Absolute protection of user privacy indirectly aids these adversaries and poses a genuine threat to public safety.
-
There is no question that an excessive concentration of power in government hands is a danger that must not be dismissed — and it has proven harmful in the past, both in the United States in the NSA affair and the Edward Snowden case, and in Israel in the Pegasus affair.
At the same time, countless lives have been saved by surveillance tools of this kind, which help identify terrorist activity and weapons every day before they can be used to harm innocent people.
The ongoing battle between privacy advocates and security forces is both necessary and healthy, because it is precisely this tension that ensures responsible use in practice. The use of artificial intelligence is no different — and one can only hope that the current struggle will lead to a thoughtful and constructive balance on the matter.
--
👋 Hi, I'm Shlomo Strauss — follow me for more content on science and technology.