ArcaMax

PARIS — Here we go again. Another “Israel First” neocon regime- change war under the pretext of a bogus “national security” risk or “freedom.” This time in Iran. Exactly what Trump’s Israeli mega-donors and benefactors have long pushed for. What a total disgrace for the FIFA Peace Prize that the world soccer association preemptively awarded him. Chalk it up to a total lack of accountability. And it’s only getting worse.

Another such glaring example emerged just in the past few days, in the lead up to this Iran invasion.

If only Clausewitz’s fog of war could be made permanently opaque, hiding the faces of those responsible for armed conflict and its results, in support of even more opaque interests. Enter Pentagon Pete Hegseth’s autonomous robot brigade.

War Secretary Hegseth crashed out online last week, ranting about Pentagon contractor Anthropic, parent company to the commercially available AI, Claude. The same tech that reportedly helped the Pentagon bomb and kill 83 people in Caracas, Venezuela, earlier this year, toss a black bag over sitting president Nicolás Maduro, and ultimately drop him in a New York prison.

Anthropic’s CEO Dario Amodei was unusually quiet after that op, despite reports that his tech worked hand-in-hand with Palantir — now infamous for its use by Israel in Gaza, and apparently so precise at targeting Hamas that an estimated 80 percent of those killed were civilians, according to a classified Israeli report.

But Amodei has now drawn a line. Any Pentagon use must include human oversight of assassinations and avoid mass surveillance of Americans.

That was enough to send Hegseth into meltdown, slapping Anthropic with the same label reserved for Chinese tech giant, Huawei: national security supply chain risk. He basically broke up with Anthropic in public, then badmouthed it to scare other Pentagon corporate suppliers from even flirting with Anthropic, lest Hegseth find out and punish them too.

Hegseth accused the company of trying to “seize veto power over the operational decisions of the United States military.” President Trump then flipped a table on social media, calling Anthropic a “Radical Left AI company run by people who have no idea what the real World is all about.” He threatened “major civil and criminal consequences” if they didn’t cooperate with a transition to OpenAI, whose founder donated $25 million to Trump’s campaign, unlike Anthropic execs.

Probably just a coincidence.

But the bigger issue is obvious. Why is the Pentagon squabbling with a supplier — in full view of the public, no less — over who gets to dictate who is spied on or unilaterally killed by a robot without human oversight? Where’s Congress with its veto power amid executive overreach, refusing funds until there’s clarity on where the buck actually stops on any mass surveillance and autonomous robot killings?

Some of us are old enough to remember when the worry was teens with a joystick in one hand, Doritos in the other, firing missiles halfway around the world. Special forces operators spoke out at the time about Obama-era centralized drone targeting inducing error and excess collateral damage, ultimately stirring the hornet’s nest in the Global War on Terror. “Obama’s drone war, a ‘recruitment tool’ for Isis, say US air force whistleblowers,” read a Guardian headline in 2015.

At least a human was pulling the trigger back then. Hard to believe that can now be considered a plus relative to what we’re dealing with nowadays. At least there was a traceable chain of command. But who is responsible for self-guided robots deciding life and death based on data alone, or for any downstream consequences. Anyone?

Foundations of accountability have been corroding for decades. Congress hasn’t asserted its exclusive right over warfare since WWII. The Korean War was merely a “police action,” President Truman said. An excuse revived by Team Trump, characterizing the invasion and regime change of Venezuela as a “law enforcement” action. The Gulf of Tonkin fiasco conveniently gave LBJ free rein in Vietnam. After that, Congress passed the War Powers Resolution, limiting deployments to 90 days without approval. Useful! Except now there are never any “troops” in war — just “advisers,” drones and now algorithmic AI robots.

Post-9/11 congressional authorizations were so effective in tethering military action to democratic accountability that it did the opposite, expanding the global war on terror to the whole planet.

Meanwhile, AI that was once confined to chatbots or recommending what socks you buy now decides who lives or dies. Algorithms assessing “threat probability” can now integrate satellite images, cellphone metadata and facial recognition — all without a human second-guessing the output. How comforting when the U.S. military has recently been shooting down its own drones on the border with Mexico.

So of course the next step in accountability regression is letting robots wage war however they see fit — against Americans or overseas. That the conscience of a single tech CEO stood between life and algorithmic death is arguably the least worrisome aspect of it all.

Just punt any blame onto AI if something goes sideways. What a comforting thought when the Pentagon had been gearing up for unprovoked regime change in Iran, contrary to everything that Trump’s base had hoped of him. Trump has already evoked the use of such tech in his and Israel’s joint invasion of Iran, referring to the use of “highly sophisticated tracking systems” in assassinating Iranian officials.

“U.S. strikes in Middle East use Anthropic, hours after Trump ban,” headlined the Wall Street Journal.

Does that also include the bombing of two schools, killing about a hundred kids? Just send the war crimes subpoenas to R2-D2 or HAL 9000, I guess.