The emerging dispute between Anthropic and the Department of Defense is more than a contract disagreement. It is a warning about what advanced AI makes possible today and what our laws currently fail to protect. According to reporting, Anthropic is asking the DOD to accept basic safety requirements that many AI researchers consider standard: independent audits of high risk systems, limits on autonomous decision making, strict controls on training data, transparency about how models are used, and prohibitions on using their AI for autonomous weapons or mass surveillance applications. These are not exotic demands; they are the kinds of guardrails many experts believe should already be in place.

A senior official recently illustrated why these safeguards matter. They noted that with modern AI, the government could “continuously monitor and analyze the public posts of every American, cross referenced against public voter registration rolls and demonstration permit records, to automatically flag civilians who live near military bases, have criticized military policy online, own firearms, and attended rallies.” That scenario is not hypothetical. It is a straightforward application of existing technology.

And it is unfolding in a country where the Constitution contains no explicit right to privacy. Every privacy protection Americans rely on today exists only because past courts inferred it from other amendments. Those interpretations have been narrowing. At the same time, the DOD is rapidly expanding its use of AI for intelligence analysis, battlefield logistics, targeting support and large scale data processing. Government capability is increasing at the exact moment the constitutional foundation for privacy is shrinking.

That is why the safeguards Anthropic is requesting deserve serious attention. But the responsibility cannot fall on private companies alone. Congress must act now. If privacy is not written into the Constitution, then Congress must establish statutory guardrails that define how federal agencies may use AI, what data they may collect, how long they may retain it, and what oversight applies. Without binding limits, any administration will inherit tools capable of monitoring Americans in ways the framers never imagined.

The question is not which company will bend. The question is whether Congress will step in before military grade AI outpaces the rights the courts are no longer willing to guarantee.

Idaho’s elected officials should not wait for next year, the next session or the next crisis. The window to set limits is now — before military grade AI grows faster than the rights.

Clinton Dunn

Dayton