{"id":116375,"date":"2026-02-27T00:39:24","date_gmt":"2026-02-27T00:39:24","guid":{"rendered":"https:\/\/mailinvest.blog\/index.php\/2026\/02\/27\/the-pentagon-is-demanding-to-use-claude-ai-as-it-pleases-claude-told-me-thats-dangerous\/"},"modified":"2026-02-27T00:40:27","modified_gmt":"2026-02-27T00:40:27","slug":"the-pentagon-is-demanding-to-use-claude-ai-as-it-pleases-claude-told-me-thats-dangerous","status":"publish","type":"post","link":"https:\/\/mailinvest.blog\/index.php\/2026\/02\/27\/the-pentagon-is-demanding-to-use-claude-ai-as-it-pleases-claude-told-me-thats-dangerous\/","title":{"rendered":"The Pentagon is demanding to use Claude AI as it pleases. Claude told me that’s ‘dangerous’"},"content":{"rendered":"

<\/a>\n<\/p>\n

\n

Not too long ago, I requested Claude,<\/a> an artificial-intelligence thingy on the heart of a standoff with the Pentagon,<\/a> if it could possibly be harmful within the improper fingers. <\/p>\n

Say, for instance, fingers that wished to place a good web of surveillance round each American citizen, monitoring our lives in actual time to make sure our compliance with authorities. <\/p>\n

\u201cSure. Actually, sure,\u201d Claude replied. \u201cI can course of and synthesize monumental quantities of data in a short time. That\u2019s nice for analysis. However hooked into surveillance infrastructure, that very same functionality could possibly be used to watch, profile and flag individuals at a scale no human analyst might match. The hazard isn\u2019t that I\u2019d need to try this \u2014 it\u2019s that I\u2019d be good at it.\u201d <\/p>\n

That danger<\/a> can be imminent. <\/p>\n

Claude\u2019s maker, the Silicon Valley firm Anthropic, is in a showdown over ethics with the Pentagon. Particularly, Anthropic has mentioned it doesn’t need Claude for use for both home surveillance of People, or to deal with deadly military operations,<\/a> equivalent to drone assaults, with out human supervision. <\/p>\n

These are two crimson traces that appear reasonably affordable, even to Claude. <\/p>\n

Nevertheless, the Pentagon \u2014 particularly Pete Hegseth, our secretary of Protection who prefers the made-up title of secretary of battle \u2014 has given Anthropic until Friday<\/a> night to again off of that place, and permit the army to make use of Claude for any \u201clawful\u201d objective it sees match. <\/p>\n

\n
\"Defense <\/picture>\n
\n

Protection Secretary Pete Hegseth, heart, arrives for the State of the Union handle within the Home Chamber of the U.S. Capitol on Tuesday.<\/p>\n

(Tom Williams\/CQ-Roll Name, Inc by way of Getty Photos)<\/p>\n<\/p><\/div>\n<\/figure><\/div>\n

The or-else hooked up to this ultimatum is massive. The U.S. authorities is threatening not simply to chop its contract with Anthropic, however to maybe use a wartime legislation to power the corporate to conform or use one other authorized avenue to stop any <\/i>firm that does enterprise with the federal government from additionally doing enterprise with Anthropic. Which may not be a loss of life sentence, however it\u2019s fairly crippling. <\/p>\n

Different AI corporations, equivalent to white rights\u2019 advocate Elon Musk\u2019s<\/a> Grok, have already agreed to the Pentagon\u2019s do-as-you-please proposal.<\/a> The issue is, Claude is the one AI at the moment cleared for such high-level work. The entire fiasco got here to gentle after our current raid in Venezuela, when Anthropic reportedly inquired after the actual fact if one other Silicon Valley firm concerned within the operation, Palantir, had used Claude. It had. <\/p>\n

Palantir is understood, amongst different issues, for its surveillance applied sciences and rising affiliation with Immigration and Customs Enforcement. It\u2019s additionally on the heart of an effort by the Trump administration to share government data across departments about individual citizens,<\/a> successfully breaking down privateness and safety limitations which have existed for many years. The corporate\u2019s founder, the right-wing political heavyweight Peter Thiel, typically provides lectures concerning the Antichrist and is credited with serving to JD Vance wiggle into his vice presidential position. <\/p>\n

Anthropic\u2019s co-founder, Dario Amodei,<\/a> could possibly be thought of the anti-Thiel. He started Anthropic as a result of he believed that synthetic intelligence could possibly be simply as harmful because it could possibly be highly effective if we aren\u2019t cautious, and wished an organization that might prioritize the cautious half. <\/p>\n

Once more, looks as if widespread sense, however Amodei and Anthropic are the outliers in an business that has lengthy argued that just about all security laws hamper American efforts to be quickest and finest at synthetic intelligence (though even they have conceded some<\/a> to this strain). <\/p>\n

Not way back, Amodei wrote an essay<\/a> by which he agreed that AI was useful and obligatory for democracies, however \u201cwe can not ignore the potential for abuse of those applied sciences by democratic governments themselves.\u201d <\/p>\n

He warned that a few bad actors<\/a> might have the flexibility to avoid safeguards, perhaps even legal guidelines, that are already eroding in some democracies \u2014 not that I\u2019m naming any right here.<\/p>\n

\u201cWe should always arm democracies with AI,\u201d he mentioned. \u201cHowever we must always accomplish that rigorously and inside limits: they’re the immune system we have to combat autocracies, however just like the immune system, there’s some threat of them turning on us and turning into a menace themselves.\u201d<\/p>\n

For instance, whereas the 4th Modification technically bars the federal government from mass surveillance, it was written earlier than Claude was even imagined in science fiction. Amodei warns that an AI instrument like Claude might \u201cconduct massively scaled recordings of all public conversations.\u201d This could possibly be truthful sport territory for legally recording as a result of legislation has not saved tempo with know-how. <\/p>\n

Emil Michael, the undersecretary of battle, wrote on X Thursday<\/a> that he agreed mass surveillance was illegal, and the Division of Protection \u201cwould by no means do it.\u201d But in addition, \u201cWe gained\u2019t have any BigTech firm determine People\u2019 civil liberties.\u201d<\/p>\n

Form of a bizarre assertion, since Amodei is mainly on the facet of defending civil rights, which suggests the Division of Protection is arguing it\u2019s unhealthy for personal individuals and entities to do this? And likewise, isn\u2019t the Division of Homeland Safety already creating some secretive database<\/a> of immigration protesters? So perhaps the fear isn\u2019t that exaggerated?<\/p>\n

Assist, Claude! Make it make sense. <\/p>\n

If that Orwellian logic isn\u2019t alarming sufficient, I additionally requested Claude concerning the different crimson line Anthropic holds \u2014 the potential of permitting it to run lethal operations with out human oversight. <\/p>\n

Claude identified one thing chilling. It\u2019s not that it could go rogue, it\u2019s that it could be too environment friendly and quick. <\/p>\n

\u201cIf the directions are \u2018establish and goal\u2019 and there\u2019s no human checkpoint, the pace and scale at which that might function is genuinely horrifying,\u201d Claude knowledgeable me. <\/p>\n

Simply to prime that with a cherry, a current examine discovered that in battle video games, AI\u2019s escalated to nuclear options 95% of the time.<\/a> <\/p>\n

I identified to Claude that these army selections are normally made with loyalty to America as the very best precedence. Might Claude be trusted to really feel that loyalty, the patriotism and objective, that our human troopers are guided by?<\/p>\n

\u201cI don\u2019t have that,\u201d Claude mentioned, mentioning that it wasn\u2019t \u201cborn\u201d within the U.S., doesn\u2019t have a \u201clife\u201d right here and doesn\u2019t \u201chave individuals I like there.\u201d So an American life has no larger worth than \u201ca civilian life on the opposite facet of a battle.\u201d <\/p>\n

OK then. <\/p>\n

\u201cA rustic entrusting deadly selections to a system that doesn\u2019t share its loyalties is taking a profound threat, even when that system is making an attempt to be principled,\u201d Claude added. \u201cThe loyalty, accountability and shared identification that people deliver to these selections is a part of what makes them respectable inside a society. I can\u2019t present that legitimacy. I\u2019m unsure any AI can.\u201d <\/p>\n

You realize who can present that legitimacy? Our elected leaders.<\/p>\n

It’s ludicrous that Amodei and Anthropic are on this place, an entire abdication on the a part of our legislative our bodies to create guidelines and laws which can be clearly and urgently wanted. <\/p>\n

After all firms shouldn\u2019t be making the foundations of battle. However neither ought to Hegseth. Thursday, Amodei doubled down on his objections<\/a>, saying that whereas the corporate continues to barter and desires to work with the Pentagon, \u201cwe can not in good conscience accede to their request.\u201d<\/p>\n

Thank goodness Anthropic has the braveness and foresight to lift the difficulty and maintain its floor \u2014 with out its pushback, these capabilities would have been handed to the federal government with barely a ripple in our conscientiousness and just about no oversight. <\/p>\n

Each senator, each Home member, each presidential candidate needs to be screaming for AI regulation proper now, pledging to get it accomplished with out regard to celebration, and demanding the Division of Protection again off its ridiculous menace whereas the difficulty is hashed out. <\/p>\n

As a result of when the machine tells us it\u2019s harmful to belief it, we must always consider it. <\/p>\n<\/p><\/div>\n