when artificial intelligence intensifies Israeli bombings

by time news

2023-12-05 19:51:41

This is undoubtedly the first time that artificial intelligence (AI) has played such an important role in such a deadly war. Intense Israeli bombardments in the Gaza Strip since the bloody Hamas attack on October 7 have already left nearly 16,000 Palestinians dead, according to the Hamas health ministry. How many of them were “recommended” as targets to the Israeli army by its “Habsora” (Gospel in English) software?

This system, including the left-wing Israeli-Palestinian site +972 and the British daily The Guardian were the first to detail the operation at the end of November, had been publicly touted by the Israeli army on its website on November 2 – even if it has since remained discreet on the subject.

While another software used by the army (“Fire factory”) makes it possible to optimize attack plans for planes and drones, “Habsora” makes target recommendations after analyzing very large amounts of data ( drone images, intercepted communications, video surveillance, etc.).

The “target factory”

The rate at which the system generates its targets is so frenetic that the Israeli military has nicknamed it the“target factory”. Several soldiers interviewed by +972, however, express doubts about their ability to make the right decisions while knowing they are “judged according to the number of targets (that they manage) to generate”.

Israel struck 15,000 targets in the first 35 days of bombing after October 7. This is three times more than during the previous deadliest war in Gaza, in 2014, which lasted 51 days.

Seven years later, during the so-called “eleven-day” war of May 2021, “Habsora” was used for the first time – the Israeli army had also said it was leading the “first AI war”. The system then generated 100 targets per day, according to former chief of staff Aviv Kochavi, interviewed at the end of June by the Israeli media YNet. “To put it in perspective, in the past we produced 50 targets in Gaza in a year. Today, this machine created 100 targets in a single day, and 50% of them were attacked. »

Is this ratio still relevant? Hard to say. The use of this technology is in any case accelerating “considerably” the decision-making of the Israeli military command, and “significantly increases the number of keystrokes”summarizes Laure de Roucy-Rochegonde, researcher at the French Institute of International Relations (Ifri).

Collateral damages

“But what explains the very high mortality of these bombings is also and above all the fact that Israel has modified the parameters of its system concerning authorized collateral damage”, adds Patrice Bouveret, director of the Armaments Observatory. According to the +972 investigation, in fact, the number of civilian deaths considered acceptable to reach a single member of Hamas would have increased from “tens” has ” hundreds ” after October 7.

“With AI, you change a parameter and it instantly modifies the resultcontinues Patrice Bouveret. But this change was decided by humans. » While the colossal scale of civilian casualties in Gaza is causing worldwide concern, knowing that these “result from machines capable of precisely anticipating them” could spark renewed indignation, considers political science researcher Laure de Roucy-Rochegonde.

Trivialization?

Along with the United States, Israel is a pioneer in the military application of technologies. The country has been developing drones in particular since the 1980s, these aerial, land and naval robots capable of automated surveillance, but also shooting. In addition to scouting, as is currently the case in Ukraine, AI is also used for logistics or military mobility.

What Patrice Bouveret fears is that the systems that Israel uses today tend to “to become commonplace” during future conflicts. “Those who develop these programs want increasingly autonomous weapons. Because AI makes it possible to distance the face-to-face between the person who kills and the person killed. »

#artificial #intelligence #intensifies #Israeli #bombings

You may also like

Leave a Comment