Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallPentagon technology chief Emil Michael said his clash with Anthropic grew out of a wider disagreement about how much freedom the military should have to use AI as weapons become more autonomous—not just one hypothetical attack. Anthropic said it would not allow Claude to be used for mass surveillance of Americans or in fully autonomous weapons without human involvement. The Pentagon wanted AI available for all lawful purposes and argued that existing law and military policy addressed those risks.
What Michael said the clash was about
In an interview aired with an Associated Press report on March 7, 2026, Michael connected the dispute to possible AI use in the future Golden Dome missile-defense program and to the Pentagon’s pursuit of greater autonomy in armed drones, underwater vehicles and other systems. He said the department needed a partner prepared to work on autonomy, adding that “someday it’ll be real” and that he needed someone who would not “wig out in the middle.” Those are Michael’s words and his description of the Pentagon’s needs, not independent evidence that Anthropic had interrupted a military operation.
That framing matters because the disagreement concerned both the uses allowed under a contract and who would decide whether a particular use crossed a line. Golden Dome and the scenarios discussed in interviews were examples in a public debate; the available reporting does not establish that one hypothetical scenario definitively caused the breakdown.
Anthropic’s red lines and the Pentagon’s position
Anthropic: no mass surveillance of Americans or fully autonomous targeting
Anthropic said it could not accept unrestricted military use of Claude without protections against mass surveillance of Americans and use in fully autonomous weapons without human involvement. CEO Dario Amodei argued that frontier AI systems were not reliable enough for fully autonomous weapons at the time. His objection was framed around current capability and oversight, not as a categorical rejection of such weapons in every possible future circumstance.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
Amodei said Anthropic “cannot in good conscience accede” to the Pentagon’s demands for unrestricted use, as quoted by the Associated Press. The phrase describes the company’s stated position; it does not resolve whether the safeguards in the proposed contract would have been sufficient.
The Pentagon: access for lawful purposes, with safeguards in existing rules
The Pentagon’s public position was that it needed AI available for all lawful purposes. It maintained that existing law and military policy addressed surveillance and human oversight. Speaking to CBS News, Michael said, “At some level, you have to trust your military to do the right thing.” That was his argument for relying on military judgment and rules, not a neutral statement that a particular contract or use was lawful.
The central governance dispute was therefore whether explicit limits in a vendor contract were necessary, or whether statutes, military directives and departmental assurances provided adequate protection. Anthropic disputed that the proposed terms preserved its safeguards; the Pentagon emphasized flexibility and dependable access.
What “autonomous weapon” means
AI used to summarize intelligence or assist operational planning is not automatically an autonomous weapon. The key distinction is whether a human selects targets, can intervene during an engagement, or is out of the loop after activation. The Congressional Research Service, summarizing Department of Defense Directive 3000.09, describes three levels:
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
| Category | Human role | What the system can do |
|---|---|---|
| Semiautonomous, or human in the loop | A human operator selects the target or targets. | The system engages the targets selected by the operator. |
| Human-supervised, or human on the loop | An operator can monitor the engagement and halt it. | The system can conduct target engagement while the operator retains a supervisory ability to intervene. |
| Fully autonomous, or human out of the loop | After activation, no further operator intervention is needed for target selection and engagement. | The system can select and engage targets without further human intervention. |
Directive 3000.09 requires commanders and operators to exercise appropriate levels of human judgment over the use of force. That does not mean a person must manually control every action: human judgment can include deciding how, when, where and why a system is employed. The distinction helps explain why Anthropic’s stated objection was specifically about fully autonomous weapons, rather than every military task involving AI.
Why the disagreement went beyond one hypothetical
- Different uses carry different risks. AI supporting analysis or planning is not the same as AI participating in selecting or engaging targets. A policy about “military AI” can obscure that operational difference.
- The parties differed on the reliability threshold. Anthropic said current frontier models were not reliable enough for fully autonomous weapons. Michael emphasized preparation for future threats and the military’s need for lawful flexibility.
- They disagreed over where safeguards should sit. Anthropic sought explicit contractual restrictions. The Pentagon argued that existing law and military policy supplied protections while preserving broader access.
- The Pentagon raised continuity concerns. Michael described a need for a dependable vendor that would not restrict or interrupt support during a future operation. That was the department’s rationale as he presented it; the reporting does not establish that Anthropic had actually interrupted an operation.
What each side said happened in negotiations
Michael’s account of private discussions is contested. In a TIME interview published March 10, 2026, he described hypotheticals involving a space-based interceptor against a hypersonic missile attack and a drone-swarm attack on a military base. He said the negotiations with Anthropic lasted three months, compared with two to three weeks for other AI labs. Those durations are Michael’s comparison, not an independently audited measure.
Anthropic’s spokesperson called parts of Michael’s account of in-person exchanges “patently false,” said Amodei did not make the statements attributed to him, and said every version of the company’s proposed contract language permitted missile defense and similar uses. TIME also reported that Anthropic disputed Michael’s account of how the talks ended and said it continued negotiating in good faith. The parties’ public accounts conflict; the reporting does not independently establish which description of the private exchanges is correct.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Timeline and status reported through October 4, 2026
| Date | Reported development |
|---|---|
| July 2025 | The Department of Defense announced contracts with Anthropic, Google, OpenAI and xAI to accelerate advanced AI adoption for national-security challenges. Each contract was worth up to $200 million, according to the Congressional Research Service’s summary; that is a ceiling, not an amount established as paid to each company. |
| February 2026 | The contract dispute became public. Anthropic said it could not accept unrestricted use without protections on mass surveillance and fully autonomous weapons. The administration directed agencies to stop using Claude and designated Anthropic a supply-chain risk, according to the Congressional Research Service’s summary. |
| March 7–10, 2026 | The Associated Press reported Michael’s comments about Golden Dome and military autonomy; TIME later published his account of negotiations and Anthropic’s response. |
| September 28, 2026 | The Associated Press reported that a federal appeals court rejected Anthropic’s challenge to the supply-chain-risk designation, allowing the Pentagon to continue removing Claude models from its systems. |
As of October 4, 2026, the September 28 Associated Press report is the latest legal and phase-out status established here; it does not establish whether there was later court action or the precise current status of Claude across Pentagon systems. The Congressional Research Service also said the Defense Department was not publicly known, at the time of its report, to be using Claude—or another frontier model—within autonomous weapon systems. That statement is about what was publicly known then, not proof that no such use existed.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
