(The author is a Reuters Breakingviews columnist. The opinions expressed are his own.)
NEW YORK - IBM is making a virtue out of necessity in ending its efforts in facial-recognition technology. The $117 billion tech firm’s new chief executive, Arvind Krishna, said in a letter to U.S. lawmakers on Monday that artificial intelligence is a powerful tool for law enforcement, but it would no longer develop or sell software based on facial recognition and questioned whether it should be used by domestic law-enforcement agencies. Companies rarely abandon potential profit for public relations wins, but it’s easier when, as in Big Blue’s case, operations are small, rivals pull ahead, and regulation threatens.
Using software to identify images is a promising field, but certain aspects are a moral quagmire. In particular, selling software that can identify people to governments – and especially authoritarian ones – will bother anyone who thinks privacy is an inherent right. Worse, facial-recognition software is only as good as the underlying assumptions and data. The result is often flawed. A study by the U.S. government’s National Institute of Standards and Technology last year found that many popular algorithms falsely identified African-American faces up to 100 times more than Caucasian ones.
IBM’s decision to use visual recognition only for detecting objects rather than faces allows it to sidestep these issues, while still tapping markets such as identifying flawed objects on manufacturing lines. Sure, the company’s decision will win it points with the public, but other factors probably carried at least as much weight in IBM’s decision.
Big tech firms ranging from Amazon.com to Apple are competing in the sector – Amazon has licensed its technology to law enforcement, while Apple promotes user privacy – as well as firms like China’s Megvii, which the U.S. government put on a human-rights blacklist last year.
It would be hard for IBM to compete in offering the technology, and there may be increased restrictions on future sales, given mass protests in the United States and the European Union’s historically greater concern over user privacy – as well as Krishna’s call for a public debate over the matter. How much current revenue it’s losing isn’t clear, but since the company won’t break it out, it’s not significant. A bit of public buffing of IBM’s reputation is therefore probably worth more both today and tomorrow.
CONTEXT NEWS
- International Business Machines Chief Executive Arvind Krishna said the company will no longer offer facial recognition technology software, in a letter sent to Congress dated June 8. Krishna said artificial intelligence is a powerful tool for law enforcement, but it must not be used to promote discrimination or racial injustice.
- IBM will no longer market, sell or update products based on the technology but will support clients, Reuters reported on June 8, citing a source. It will not develop, create, research or sell facial recognition products, and its visual technology will be limited to visual object detection, not for facial analysis and identification, said the source.
(The author is a Reuters Breakingviews columnist. The opinions expressed are his own.)
(Editing by Antony Currie and Amanda Gomez) ((robert.cyran@thomsonreuters.com; Reuters Messaging: robert.cyran.thomsonreuters.com@reuters.net))
