Another knock against the robots.
Earlier this week, an Asian New Zealander posted a screenshot to Facebook showing his passport photo being rejected for having closed eyes, even though they are visibly open. Twenty-two-year-old Richard Lee—an engineer student from New Zealand who was born in Taiwan and is currently studying in Melbourne, Australia—was trying to renew his passport online using an algorithm-based photo-checker operated by New Zealand’s department of internal affairs.
Lee took the mess-up in stride. “No hard feelings on my part,” he told Reuters. “I’ve always had very small eyes and facial recognition technology is relatively new and unsophisticated.” He also told Mashable that he ended up taking more photos, one of which was accepted by the system.
Passport-photo errors are common—as many as 20% of photos submitted to the system Lee used are rejected, most often for closed eyes—but the complexities are indicative of broader issues with facial-recognition technology. As far back as 2010, Time magazine reported that people using digital cameras equipped with similar algorithms had similar issues: Cameras would warn that people’s eyes (particularly people of Asian descent) were closed when they weren’t.
The racist stereotypes surrounding Asians’ eyes goes back even further. While Lee’s passport kerfuffle may have been an accident, it’s one we don’t want algorithms repeating.