Historian uses AI to help identify Nazi in notorious Holocaust murder image (www.theguardian.com)

German historian Jürgen Matthäus announced in a peer-reviewed paper that he has likely identified the Nazi gunman in a notorious Holocaust photograph—taken on 28 July 1941 in the citadel of Berdychiv, Ukraine—as Jakobus Onnen of Einsatzgruppe C. The finding combines decades of traditional archival sleuthing, eyewitness trails and collaboration with Bellingcat volunteers, who applied AI-powered facial comparison to family photos of Onnen. Technical experts described the algorithmic match as “unusually high,” though Matthäus cautions that historical prints lower confidence compared with contemporary forensic matches (rarely reaching the 98–99.9% levels seen today), so the AI result is presented alongside extensive circumstantial documentary evidence rather than as a lone proof. The case is significant because it demonstrates how AI and open-source forensic techniques can assist qualitative historical inquiry—helping attribute individual responsibility in mass atrocities where many perpetrators and victims remain nameless. Matthäus frames the tools as amplifying, not replacing, traditional methods: human judgment, archival context, and corroborating records remain essential. The project also continues to seek the identity of the kneeling victim using Soviet-era community records and might use AI again if comparable portraits are found, underscoring both the potential and the limits of digital methods in reconstructing contested, century-old crimes.