How US Intelligence and an American Company Feed Israel’s Killing Machine in Gaza

Now, with Israel’s ongoing war in Gaza, critical information from NSA continues to be used by Unit 8200, according to a number of sources, to target tens of thousands of Palestinians for death—often with US-supplied 2,000-pound bombs and other weapons. And it is extremely powerful data-mining software, such as that from Palantir, that helps the IDF to select targets. While the company does not disclose operational details, some indications of the power and speed of its AI can be understood by examining its activities on behalf of another client at war: Ukraine. Palantir is “responsible for most of the targeting in Ukraine,” according to Karp. “From the moment the algorithms set to work detecting their targets [i.e., people] until these targets are prosecuted [i.e., killed]—a term of art in the field—no more than two or three minutes elapse,” noted Bruno Macaes, a former senior Portuguese official who was given a tour of Palantir’s London headquarters last year. “In the old world, the process might take six hours.”

The company is currently developing an even more powerful AI targeting system called TITAN (for “Tactical Intelligence Targeting Access Node”). According to Palantir, TITAN is a “next-generation Intelligence, Surveillance, and Reconnaissance ground station enabled by Artificial Intelligence and Machine Learning to process data received from Space, High Altitude, Aerial and Terrestrial layers.” Although designed for use by the US Army, it’s possible that the company could test prototypes against Palestinians in Gaza. “How precise and accurate can you know a system is going to be unless it’s already been trained and tested on people?” said Catherine Connolly of the Stop Killer Robot coalition, which includes Human Rights Watch and Amnesty International.

The most in-depth examination of the connection between AI and the massive numbers of innocent Palestinian men, women, and children slaughtered in Gaza by Israel comes from an investigation recently published by +972 Magazine and Local Call. Although Palantir is not mentioned by name, the AI systems discussed by the journalists appear to fit into the same category. According to the lengthy investigation, Unit 8200 is currently using a system called “Lavender” to target thousands of alleged Hamas fighters. But the magazine also reported that, while before the brutal attack on October 7 Israel’s rules of engagement tightly restricted the numbers of non-combatant casualties allowed in targeting a single alleged Hamas militant, such limitations have been loosened in the months since to the point of allowing for the slaughter of dozens of Palestinian noncombatants (including women and children) for each targeted individual. President Joe Biden’s December warning that Israel was losing international support because of its “indiscriminate bombing” of Gaza appears to have had no effect.

The +972 Magazine report details how the Israeli military uses powerful algorithms to sort through enormous volumes of surveillance data—phone, text, and digital——to come up with lengthy kill lists of targets. And adding to that haul would be the data from the NSA intercepts of Palestinians in the United States communicating with their families in Gaza—a process that continued after Snowden left NSA, according to a number of sources.

In 2014—more than a year after Snowden turned up in Moscow—Unit 8200’s targeting of innocent Palestinian in the occupied territories was so extreme that it even caused 43 veterans of the unit, including many serving in the reserves, to go public and accuse the organization of startling abuses. They declared that they had a “moral duty” to no longer “take part in the state’s actions against Palestinians.”

In a letter to their commanders, Prime Minister Benjamin Netanyahu, and the head of the Israeli army, they charged that Israel used the information collected against innocent Palestinians for “political persecution.” And in testimonies and interviews given to the media, they specified that data were gathered on Palestinians’ sexual orientations, infidelities, money problems, family medical conditions and other private matters that could be used to coerce Palestinians into becoming collaborators or create divisions in their society.”

Several years ago, Brig. Gen. Yossi Sariel, the current director of Unit 8200, published a book outlining a supposedly fictional and far-reaching AI system. But the journalists from +972 and Local Call discovered that the super-powerful target generation machine he wrote about then as fiction actually exists. “During the first weeks of the war, the army almost completely relied on Lavender,” they write, “which clocked as many as 37,000 Palestinians as suspected militants—and their homes—for possible air strikes.” And from the beginning, there was little attempt to verify or justify the thousands of names generated by the machine’s “kill list.”

“One source,” they write, “stated that human personnel often served only as a ‘rubber stamp’ for the machine’s decisions, adding that, normally, they would personally devote only about ‘20 seconds’ to each target before authorizing a bombing—just to make sure the Lavender-marked target is male. This was despite knowing that the system makes what are regarded as ‘errors’ in approximately 10 percent of cases and is known to occasionally mark individuals who have merely a loose connection to militant groups, or no connection at all.”

But it gets even worse. They found that the Israeli army deliberately and systematically bombed the homes of targeted individuals—killing entire families—simply because another AI algorithm told them the individual was there. “The result,” they write, “as the sources testified, is that thousands of Palestinians—most of them women and children or people who were not involved in the fighting—were wiped out by Israeli airstrikes, especially during the first weeks of the war, because of the AI program’s decisions.”

Such actions were likely contributing factors in the recent decision by the United Nations Human Rights Council to adopt a resolution calling for Israel to be held accountable for possible war crimes and crimes against humanity committed in Gaza. Twenty-eight countries voted in favor, with only six voting against the resolution. Still, in the White House and on Capitol Hill, there seems little concern for the dangers of Israel’s deadly weaponization of AI and the technology’s connection to the staggering numbers of innocent men, women, children, and whole families massacred in Gaza. Washington director for Human Rights Watch Sarah Yager told Politico, “Nobody has any insight including, I would say, U.S. policymakers on how Israel is conducting this war.”

For years, the United States has had strict regulations on the export of weapon systems to foreign countries because of the lack of accountability once in the users’ possession, and the potential for serious war crimes. Even Palantir CEO Alex Karp has argued that “the power of advanced algorithmic warfare systems is now so great that it equates to having tactical nuclear weapons against an adversary with only conventional ones.” Israel’s indiscriminate killing in Gaza offers the perfect example of why it’s time to also begin far stricter regulation of the export of AI systems, like those developed by Palantir. Systems that, as Karp suggested, are the digital equivalent of a weapon of mass destruction. After all, it’s not just the bomb that kills, but the list that puts you and your family under it.