Responsible use of facial biometrics in focus at World Economic Forum
As issues related to artificial intelligence, and facial biometrics in particular are receiving significantly increasing attention from the public and elected representatives, several presentations at The World Economic Forum’s (WEF’s) annual meeting for 2019 in Davos, Switzerland have explored the potential benefits and risks posed by these advanced technologies.
Microsoft CEO Satya Nadella said in a presentation that self-regulation by facial recognition providers is important, not sufficient on its own, CNBC reports.
“One of the things that I feel today is, in the marketplace, there’s competition; there’s no discrimination between the right use and the wrong use of facial recognition,” Nadella says.
Microsoft is preparing to put its own principles for self-regulation, as outlined by company President Brad Smith, into operation.
“At the same time we also welcome any regulation that helps the marketplace not be a race to the bottom,” which Nadella says could ultimately lead to more heavy-handed regulation. Nadella called GDPR a “fantastic start,” but suggested that global regulation is needed.
An easy fix for bias
MIT researcher Joy Buolamwini told an audience that after her research on discrepancies in accuracy between facial biometric systems matching subjects with different skin color was sent to IBM, the company dramatically improved its algorithm. In new testing, Buolamwini found accuracy for dark-skinned males improved from 88 percent to 99.4 percent, and for dark-skinned females improved from 65.3 percent to 83.5 percent. Accuracy for light-skinned females improved 92.9 percent to 97.6 percent, while for lighter-skinned males it remained at 97 percent.
“So for everybody who watched my TED Talk and said, ‘Isn’t the reason you weren’t detected because of, you know, physics? Your skin reflectance, contrast, et cetera,’ — the laws of physics did not change between Dec. 2017, when I did the study, and 2018, when they launched the new results,” Buolamwini said. “What did change is they made it a priority.”
She also noted that no matter how effective technology is, usage decisions made by people can still produce unwanted outcomes, as when the New York City Police Department used IBM facial recognition for years without disclosing it. This is why Buolamwini calls on businesses to take the Safe Face Pledge to use facial recognition responsibly.
Yahoo Finance reports that the WEF and LinkedIn’s Global Gender Gap Report 2018 shows that only 22 percent of people working on AI are women, and LinkedIn Co-founder and VP of Products Allen Blue suggested that the biases observed in facial recognition systems stem from the lack of diversity in the engineering teams that develop them.
Heavyweight panel weighs facial recognition benefits and risks
A heavyweight panel discussion specific to facial recognition was hosted by Henry Blodget, the CEO, co-founder, and editorial director of Insider Inc., which publishes Business Insider. World Economic Forum Centre for the Fourth Industrial Revolution Head of Artificial Intelligence and Machine Learning Kay Firth-Butterfield, Human Rights Watch Executive Director Kenneth Roth, Member of European Parliament Marietje Schaake, and Microsoft’s Brad Smith participated as panelists.
Roth began his remarks by acknowledging the value of facial recognition for a variety of use cases, before suggesting that regulation must be put in place before major potential negative impacts occur, in part because doing so after could be more difficult. He then pointed out that even in public, people can assume a certain degree or type of anonymity, in that tracking an individual in a crowded urban setting with numerous people around is highly impractical. The law does not generally recognize privacy rights in public, but Roth suggests that it may have to evolve to do so.
Shaake pointed out that the assumption Roth refers to is already unsafe due to the level of proliferation facial recognition technology has reached, and that the cross-use of data needs to be considered as part of the same issue.
“While I think it’s important that we have this discussion, we shouldn’t pretend that we can be ahead of the curve,” Shaake argues. “It’s already out there, and I wish we would have had this discussion much earlier.”
Firth-Butterfield noted that in her work on human rights, she has seen major benefits from facial recognition technology in fighting human trafficking. At the same time, she expressed particular concern with the use of AI and facial recognition in children’s toys, and pointed out that such use cases bring concerns about the technology from the public to the private space.
Microsoft continues to receive kudos for its proactive stance, but the fight over regulation of facial recognition, and potentially other biometric or AI applications, is just beginning.
artificial intelligence | best practices | biometrics | ethics | facial recognition | Joy Buolamwini | Microsoft | privacy