Face recognition glitch

Faulty facial recognition caused Alonzo Sawyer, a U.S. citizen, to spend nine days in jail even though he was innocent. This is not an isolated case in the US.

Arrest based on AI facial software.

Police charged Sawyer based on AI facial software that allegedly identified him on surveillance video. Despite his wife's testimony that he was sitting on the couch at the time of the crime and the alleged evidence video showing the suspect was not Sawyer, he was arrested.

Conspicuous features ignored

Sawyer's wife attempted to prove his innocence. In addition to her testimony of the evening spent on the couch, she also recognized from the alleged evidence video that the suspect there could in no way be her husband. That's because the person who had robbed the bus was significantly smaller and younger, had no beard and no gaps between his teeth - unlike her husband. As Wired reports, Sawyer had a noticeable gait, according to his wife. She also said she was unable to see that in the video footage.

Software often faulty in black people

It is well known that appropriate software has a higher error rate in non-white people. Still, many U.S. law enforcement agencies rely on facial recognition software such as the controversial Clearview AI and ignore noticeable discrepancies such as in body type.

Unregulated use of technology in the U.S.

Although some U.S. states have already banned or restricted the technology, its use is largely unregulated in the United States. Cases like this are sure to create further conversation.

Luck of the draw

Sawyer was ultimately lucky in that a victim of the robbery identified another suspect who was in fact the perpetrator.