Debra Jean Beasley is a knockout, a sun-kissed blonde with a great smile. She’s also a registered sex offender, according to CreepShield, which finds the face in a provided test photo 99 percent similar to Beasley’s mug shots. Copy an image URL into its search bar and, boom, decide if possible matches are close enough to cancel that coffee date.
The new facial recognition search engine, created by local tech developer Kevin Alan Tussy, aims to make online dating safer by allowing users to check headshots against a database of over 475,000 sex offenders (cached from publicly accessible federal and state registries). But scans of “facial characteristics” don’t make “special considerations” for gender, race or age, explaining why CreepShield found the test blonde 48 percent similar to an African American woman and 46 percent similar to a 69-year-old mustachioed man. Ultimately, the site can’t say if a photo is an exact match, and it doesn’t claim to have all offenders on file, though Tussy is cultivating more data. In a news release, he said: “While this technology can’t identify everyone out there with bad intentions, it’s a huge step in the right direction.”