3 October 7 perpetrators identified with facial recognition software
On the morning of October 7, 2023, a bloodied Joshua Mollel raised his hands in surrender as armed Palestinians dragged him from Kibbutz Nahal Oz. Shortly thereafter, Mollel was murdered, and his body was taken across the border to Gaza.
More than two years later, Mollelâs killers and the those who murdered many other victims during the Hamas-led massacre have not been identified. But facial recognition tools can see what was not visible before.
An estimated 7,000 Palestinians participated in the October 7 attack. Since that day, the Israel Defense Forces (IDF) has published the names of numerous individuals who carried out attacks and abductions, but more identities remain unknown. The terroristsâ own recording devices are important sources of evidence, and the available footage is extensive, but only a small portion lends itself to identifying the perpetrators. However, the combined use of facial recognition tools and social media exploitation can yield verifiable identity data.
FDDâs Long War Journal investigated the identities of suspects involved in the Hamas-led October 7 attack in southern Israel, and identified three individualsâAmjad Qadoum, Yasir Samir, and Mustafa Shaheenâwho were involved in the atrocities committed that day.
Amjad Qadoum
On December 17, 2023, footage depicting Mollelâs capture and killing was uploaded to the social media platform Telegram and shared widely. Mollelâs father expressed anger at both his sonâs killers and those who made the video public.
The footage shows several armed men holding and speaking to a bloodied Mollel. Seconds into the video, a man wearing a wide-brimmed hat and holding an object with a black handle enters the frame. A second clip shows the immediate aftermath of Mollelâs stabbing, with Mollel on the ground and bleeding from wounds to his torso. The man with the wide-brimmed hat is holding a bloody knife and standing near Mollelâs body.
On its Telegram account, the Fatah-affiliated Al Aqsa Martyrsâ Brigades posted another video of Mollelâs lifeless body, titled âLive scenes from the battlefield.â The footage also shows the man with a wide-brimmed hat, knife in hand, and matching clothes.
Long War Journal used two artificial intelligence (AI) facial recognition tools to search for Mollelâs killer: Lenso.ai and PimEyes. Both platforms returned an open-source image from a Special Broadcasting Service (SBS) reproduction of an August 6, 2022, Agence France-Presse article. The photoâpublished by SIPA Images and credited to Mahmoud Issaâdepicted unnamed individuals mourning the death of a family member named Alaa Qadoum.
A Facebook search for the family name of Qadoum in Arabic (ŮŘŻŮŮ
) turned up a profile of Amjad Qadoum, whose facial features closely resemble the suspect in the Mollel videos.
LWJ used Amazon Rekognitionâs âFace Comparisonâ tool to assess whether the image in the Amjad Qadoum Facebook profile matched the face of Mollelâs murderer. Amazon Rekognition assessed a 99.4 percent similarity.
It is unclear if the name Amjad Qadoum is authentic or an alias. However, the suspectâs link to Alaa Qadoum indicates that the name is likely authentic. Open-source information does not indicate whether Amjad Qadoum is still alive. Nor is it clear whether he was an official member of a terrorist organization.
On November 5, 2025, Hamas transferred the body of Joshua Mollel to Israel as part of a ceasefire agreement.
Yasir Samir
Open-source photographic evidence circulated on social media depicts several Palestinians forcibly abducting Raz Ben Ami from Kibbutz Beâeri on October 7. Among the suspects in the video, LWJ identified Yasir Samir as one of the suspects involved in the kidnapping.
The original source of the photograph is unknown, but multiple independent sources indicate that it is authentic. Channel 12 News Israel broadcast the photo in a report detailing the kidnapping of Raz and her husband from Kibbutz Beâeri on October 7. Separately, Israelâs public broadcaster, Kan News, presented the same photo relating to the kidnapping of the couple.
Additionally, Kan broadcast a separate video of the Ben Ami kidnapping. The victim and the suspects in the footage are wearing the same clothes seen in the photograph shown by Kan and Channel 12. While the photo and video were clearly taken by a participant in or an observer of the kidnapping, it is not clear how Israeli journalists obtained the photo, which provides an unobstructed view of the face of one of the kidnappers.
LWJ queried Lenso.ai and PimEyes and found multiple matches for the man in this photo on Gaza-based websites (shown below). Lenso.ai returned an article on the Nuseirat Municipalityâs official website (nuseirat.ps) that shows a gathering of individuals to support the Nuseirat camp in central Gaza. PimEyes returned an article from Alkofiya.tv, an online Palestinian news organization, showing an individual who resembled the suspect performing at a music event with a band. However, neither Lenso.ai nor PimEyes identified the suspect by name.
To obtain additional information on the suspect (circled in white in the image above), LWJ used PimEyes on all the individuals in the Nuseirat.ps photo. The query returned a link to a Facebook page under the name of Mohammed al Hour (circled in red in the image above). On Hourâs Facebook page, LWJ located a photo of an individual who resembled the suspect receiving a haircut. The photograph was tagged to the account of Yasir Samir, who matched the suspectâs appearance from the Raz Ben Ami kidnapping photograph.
Amazon Rekognition compared facial features in photos from Samirâs Facebook page with those of the Ben Ami kidnapper. The query returned a 99.9 percent similarity, verifying that Samir was part of the attack on October 7. Open-source evidence provides no clear indication that he belonged to a specific terrorist group, nor could it be determined if the name Yasir Samir is authentic or an alias.
Hamas released hostage Ben Ami on November 29, 2023, as part of a temporary ceasefire agreement with Israel.
Mustafa Shaheen
A photograph by Ali Mahmud, an Associated Press photographer, served as a reference image for LWJ to identify a group of armed suspects in the back of a pickup truck with the body of German-Israeli national Shani Louk during the October 7 massacre. Following an investigation, LWJ identified one of the gunmen as Mustafa Shaheen.
Hamas abducted Louk from the Nova music festival and murdered her. While the Israeli media outlet ILTV reported that airstrikes killed the suspect days after the abduction, neither the outlet nor the IDF published his identity.
LWJ queried Lenso.ai using the Associated Press photo, and the tool returned an article on Airwars.org that provided details about an Israeli strike on a home belonging to the Shaheen family in Jabalia. The website provided a picture profile of Mustafa Shaheen and details about the Israeli airstrike that killed him and members of his family on October 10, 2023.
Amazon Rekognitionâs Face Comparison found a 99.9 percent similarity between the reference image from October 7 and the profile picture on Airwars.org.
Mustafa Shaheen is likely the suspectâs true name. LWJ found a Facebook page called âMustafa Shaheen Glass and Mirrors Workshopâ that contains photos and a video of an individual who appears to be Shaheen.
The IDF recovered Loukâs body in Gaza on May 16, 2024.
Conclusion
Evidence collected and processed with facial recognition tools identified Amjad Qadoum, Yasir Samir, and Mustafa Shaheen as participants in the October 7 massacre. These three names contribute to the ongoing investigation seeking to identify the thousands of perpetrators of the October 7 attacks. There are no publicly available Israeli government figures specifying how many suspects in the attack remain to be identified, but facial recognition software and open-source analysis can lead to naming other perpetrators.
A version of this article was previously published at FDD.org.
How it works
Once you click Generate, Ollama reads this article and crafts 5 comprehension questions. Your answers are graded against the article content â general knowledge won't be enough. Score 70+ to count toward your certificate.
Questions are cached â you'll always get the same 5 for this article.