In the wake of the devastating Hamas terrorist attacks that left 1,300 dead and hundreds missing in Israel, a group of cyber experts initiated a groundbreaking volunteer project to identify and locate the missing individuals. Led by Refael Franco, former deputy director of the Israeli government’s National Cyber Directorate and current co-founder of security startup Code Blue Cyber, the team developed a code to scour social media platforms such as Telegram, Twitter, and TikTok for images and videos related to the attacks. Over a span of two weeks, the project successfully identified around 60 missing people and provided valuable leads on the status and whereabouts of five others.

The project, conducted in what can be described as a “cyber war room,” began by creating a map of locations where Hamas was believed to be operating. This strategic map was based on the analysis of videos made by Hamas of captured Israelis, with team members drawing on their previous in-the-field experience around Gaza. Subsequently, the team sought out relevant social media accounts associated with these locations, including neighborhood Facebook groups and pages sharing war footage, and began scraping images and videos.

While some social media platforms, like Facebook, required manual searches due to restrictions on scraping, Telegram proved to be a fruitful source of information and could be scraped automatically. One of the project’s key collaborators was the Israeli startup Tagbox, which typically utilizes artificial intelligence algorithms to sort and categorize images and videos for businesses. Tagbox temporarily shifted its focus to assist Franco’s effort, helping filter the vast amount of scraped images for relevance before subjecting them to Amazon’s Rekognition algorithm.

According to Guy Barner, C.E.O. and co-founder of Tagbox, the project faced the challenge of dealing with close to 1,000 individuals for whom there was no status, potentially being alive or dead in Israel or Gaza. The team meticulously analyzed tens of thousands of photos and video files, with Amazon’s Rekognition algorithm proving instrumental in identifying obscured faces that human analysts might have missed. Although the initiative has not yet resulted in actual rescues, it has shown promise, prompting the Israeli Defense Forces (I.D.F.) to integrate it into their ongoing rescue operations.

Franco emphasized the sensitivity of the data collected during the project, leading to its transition from a volunteer effort to an official operation under the Israeli government. As per the I.D.F.’s estimates, up to 200 Israelis are either missing or have been kidnapped by Hamas.

This deployment of facial recognition technology is the second documented use case in the Israel-Hamas conflict. Corsight A.I., another Israeli company, was reported to use facial recognition in hospitals to check inpatients against a database of missing individuals.

Amazon’s Rekognition tool, introduced in 2016, initially gained attention for its cost-effectiveness and user-friendly facial recognition capabilities. However, controversies arose when its inaccuracies were exposed. Critics, including the American Civil Liberties Union, raised concerns about false matches, particularly affecting people of color. Amazon responded by halting sales of the tool to police departments over ethical concerns.

As the Israeli cyber experts continue to utilize advanced technology in their humanitarian efforts, the intersection of artificial intelligence and crisis response takes center stage. The success of this initiative not only highlights the potential of A.I. in locating missing persons but also raises ethical considerations surrounding the use of facial recognition technology in sensitive contexts.