Computer Fail: Facial Recognition Software Rejects Asian’s Passport Photo Due To 'Closed Eyes,' Sparks Debate On Computers And Racism

This incident has gone viral and has sparked a heated debate on computers and racism. A passport photo of an Asian man was rejected by the facial recognition software of the New Zealand government site. The passport photo was rejected due to a "Subject eyes are closed" error message when clearly the Asian man's eyes are open.

Richard Lee, a Kiwi of Asian descent, was trying to renew his passport when the New Zealand government website rejected his passport photo. Mr. Lee says he received a "Subject eyes are closed" message on his photo on which his eyes are clearly open.

Richard Lee, currently in Australia studying Aerospace Engineering and Business Management, said he had to get his picture taken again and eventually one of them got accepted. He has conducted New Zealand's Internal Affairs and was informed that his first photo was rejected due to uneven lighting on the face.

The 22-year old, also known as DJ Richy Fancy, posted the incident on Facebook where lots of his friends started sharing the photo online, making it viral. Most of the people who saw the post are saying how even computers are becoming racists. Also, others were chiming in that they too experienced this kind of error with their passport photos.

However, for Richard Lee, he just shared the post because he found it hilarious. When asked about the incident, Mr. Lee said he was not bothered by it and did not think it was a race thing that had his passport photo rejected. He adds, "It was a robot. No hard feelings," in an interview with BBC.

According to Reuters, a spokesperson for the New Zealand's Department of Immigration says 20 percent of people's passport photo were rejected due to the subject's closed eyes and adds that the error message was pretty generic. In addition, the Department of Immigration does defend its facial recognition software as one of the most technologically advanced software in the world. But errors like Richard Lee has experienced are still common.