The Facial Recognition Industry Promises to Regulate Itself. Sure, Okay.
Members of the industry group proposing ethics guidelines have broken those guidelines already
Welcome to General Intelligence, OneZero’s weekly dive into the A.I. news and research that matters.
While American cities and Congress weigh whether facial recognition technology has a place in the U.S., the world’s largest facial recognition companies are trying to get out ahead of legislation by publishing a new set of ethics guidelines.
This week, the Security Industry Association, a trade organization representing the world’s largest security firms, like NEC, Idemia, Dahua, and Hikvision, published broad principles for what they consider to be responsible use of facial recognition. The document also warns against broad regulation of facial recognition technology. And banning the industry outright, SIA suggests, is the wrong decision, too.
“SIA believes all technology products, including facial recognition, must only be used for purposes that are lawful, ethical, and nondiscriminatory,” the document says. “Blanket moratoriums and bans shutter both proven current uses and potential future benefits across many uses.”
The SIA’s principles sound great on paper: accuracy, nondiscrimination, transparency, accountability, and resisting the technology’s use for mass surveillance by law enforcement.
In reality, SIA’s members seem indifferent to whether these principles are upheld. The organization’s most prominent members sell real-time facial recognition to law enforcement. NEC’s installation of real-time facial recognition in the U.K. was shown to be wildly ineffective in independent tests and was actually ruled by the U.K. Court of Appeal as a breach of human rights and data privacy laws.
Nearly every recommendation of the SIA principles for using facial recognition responsibly has been violated by either the companies selling the technology or the customers they sold it to.