By Sheera Frenkel
Within minutes of walking through an Israeli military checkpoint along the Gaza Strip’s central highway Nov. 19, Palestinian poet Mosab Abu Toha was asked to step out of the crowd. He put down his 3-year-old son, whom he was carrying, and sat in front of a military jeep.
Half an hour later, Abu Toha heard his name called. Then he was blindfolded and led away for interrogation.
“I had no idea what was happening or how they could suddenly know my full legal name,” said the 31-year-old, who added that he had no ties to the militant group Hamas and had been trying to leave Gaza for Egypt.
It turned out Abu Toha had walked into the range of cameras embedded with facial recognition technology, according to three Israeli intelligence officials who spoke on the condition of anonymity. After his face was scanned and he was identified, an artificial intelligence program found that the poet was on an Israeli list of wanted persons, they said.
Abu Toha is one of hundreds of Palestinians who have been picked out by a previously undisclosed Israeli facial recognition program that was started in Gaza late last year. The expansive and experimental effort is being used to conduct mass surveillance there, collecting and cataloging the faces of Palestinians without their knowledge or consent, according to Israeli intelligence officers, military officials and soldiers.
The technology was initially used in Gaza to search for Israelis who were taken hostage by Hamas during the Oct. 7 cross-border raids, the intelligence officials said. After Israel embarked on a ground offensive in Gaza, it increasingly turned to the program to root out anyone with ties to Hamas or other militant groups. At times, the technology wrongly flagged civilians as wanted Hamas militants, one officer said.
The facial recognition program, which is run by Israel’s military intelligence unit, including the cyberintelligence division Unit 8200, relies on technology from Corsight, a private Israeli company, four intelligence officers said. It also uses Google Photos, they said. Combined, the technologies enable Israel to pick faces out of crowds and grainy drone footage.
Three of the people with knowledge of the program said they were speaking out because of concerns that it was a misuse of time and resources by Israel.
An Israeli army spokesperson declined to comment on activity in Gaza but said the military “carries out necessary security and intelligence operations, while making significant efforts to minimize harm to the uninvolved population.” He added, “Naturally, we cannot refer to operational and intelligence capabilities in this context.”
Facial recognition technology has spread across the globe in recent years, fueled by increasingly sophisticated AI systems. While some countries use the technology to make air travel easier, China and Russia have deployed the technology against minority groups and to suppress dissent. Israel’s use of facial recognition in Gaza stands out as an application of the technology in a war.
Matt Mahmoudi, a researcher with Amnesty International, said Israel’s use of facial recognition was a concern because it could lead to “a complete dehumanization of Palestinians” where they were not seen as individuals. He added that Israeli soldiers were unlikely to question the technology when it identified a person as being part of a militant group, even though the technology makes mistakes.
Israel previously used facial recognition in the West Bank and east Jerusalem, according to an Amnesty report last year, but the effort in Gaza goes further.
In the West Bank and east Jerusalem, Israelis have a homegrown facial recognition system called Blue Wolf, according to the Amnesty report. At checkpoints in West Bank cities such as Hebron, Palestinians are scanned by high-resolution cameras before being permitted to pass. Soldiers also use smartphone apps to scan the faces of Palestinians and add them to a database, the report said.
In Gaza, which Israel withdrew from in 2005, no facial recognition technology was present. Surveillance of Hamas in Gaza was instead conducted by tapping phone lines, interrogating Palestinian prisoners, harvesting drone footage, getting access to private social media accounts and hacking into telecommunications systems, Israeli intelligence officers said.
After Oct. 7, Israeli intelligence officers in Unit 8200 turned to that surveillance for information on the Hamas militants who breached Israel’s borders. The unit also combed through footage of the attacks from security cameras, as well as videos uploaded by Hamas on social media, one officer said. He said the unit had been told to create a “hit list” of Hamas members who participated in the attack.
Corsight was then brought in to create a facial recognition program in Gaza, three Israeli intelligence officers said.
The company, with headquarters in Tel Aviv, says on its website that its technology requires less than 50% of a face to be visible for accurate recognition. Robert Watts, Corsight’s president, posted this month on LinkedIn that the facial recognition technology could work with “extreme angles, (even from drones,) darkness, poor quality.”
Corsight declined to comment.
To supplement Corsight’s technology, Israeli officers used Google Photos, the free photo sharing and storage service from Google, three intelligence officers said. By uploading a database of known persons to Google Photos, Israeli officers could use the service’s photo search function to identify people.
Google’s ability to match faces and identify people even with only a small portion of their face visible was superior to other technology, one officer said. The military continued to use Corsight because it was customizable, the officers said.
A Google spokesperson said Google Photos was a free consumer product that “does not provide identities for unknown people in photographs.”
The facial recognition program in Gaza grew as Israel expanded its military offensive there. Israeli soldiers entering Gaza were given cameras equipped with the technology. Soldiers also set up checkpoints along major roads that Palestinians were using to flee areas of heavy fighting, with cameras that scanned faces.
The program’s goals were to search for Israeli hostages, as well as Hamas fighters who could be detained for questioning, the Israeli intelligence officers said.
The guidelines of whom to stop were intentionally broad, one said. Palestinian prisoners were asked to name people from their communities who they believed were part of Hamas. Israel would then search for those people, hoping they would yield more intelligence.
Abu Toha, the Palestinian poet, was named as a Hamas operative by someone in the northern Gaza town of Beit Lahia, where he lived with his family, the Israeli intelligence officers said. The officers said there was no specific intelligence attached to his file explaining a connection to Hamas.
Abu Toha said he was beaten and interrogated in an Israeli detention center for two days before being returned to Gaza with no explanation. He wrote about his experience in The New Yorker, where he is a contributor. He credited his release to a campaign led by journalists at The New Yorker and other publications.
Upon his release, Israeli soldiers told him his interrogation had been a “mistake,” he said.
In a statement at the time, the Israeli military said Abu Toha was taken for questioning because of “intelligence indicating a number of interactions between several civilians and terror organizations inside the Gaza Strip.”