During Facebook CEO Mark Zuckerberg’s testimony to members of the U.S. Senate on Tuesday, the main topic of discussion was regulation.
Zuckerberg was asked why Facebook should be trusted to self-regulate, what kind of regulation he would like to see, whether he would be willing to endorse the CONSENT Act (which would allow the FTC to regulate data privacy protections for personal data), whether he still endorses the Honest Ads Act, and so on.
One of the most telling moments in conversations about regulation came from Senator Orrin Hatch of Utah, who, after mentioning the potential danger of excessive government regulation, asked what kind of regulation would not help solve the kinds of problems made clear by the Cambridge Analytica trainwreck. Zuckerberg responded by saying that a crackdown limiting facial recognition could stifle innovation.
“I think that there’s a balance that’s extremely important to strike here, where you obtain special consent for special features like face recognition,” Zuckerberg said. “But we still need to make it so that American companies can innovate in these areas, or else we’re going to fall behind Chinese competitors and others around the world who have different regimes for different new features like that.”
In two weeks, at its annual developer conference F8, Facebook was expected to debut its first hardware device for consumers driven by voice control and facial recognition, but that roll-out has reportedly been delayed by privacy concerns.
Zuckerberg’s comments about facial recognition come days after Chinese facial recognition and computer vision startup SenseTime raised $600 million in a round led by Alibaba. IDG Group and Qualcomm are also investors. At a $3 billion valuation, the three-year-old company is now the most valuable AI startup in the world.
Also based in China, Megvii Face++ raised $460 million in the fall of 2017 in a round led by an investment fund backed by the Russian and Chinese governments.
Players like Face++ and SenseTime partner with the Chinese government, which has made a facial recognition database of its 1.3 billion citizens that powers things like CCTV surveillance and smart glasses for police to identify people in seconds. Partnership with government is giving Chinese firms the power to outpace companies like Facebook.
For example, last year Face++ outperformed Google, Facebook, and Microsoft in a bout at International Conference on Computer Vision.
Future Today Institute founder Amy Webb recently noted that if data is the new oil in the age of AI, China controls the largest repository of human data on the planet. Add a government with little regard for human rights, and Webb argues that China will surpass the rest of the world in AI by the end of 2018 — not by 2030 as China has projected.
Facebook’s ‘arms race’ against Russia
Zuckerberg’s comments regarding Chinese facial recognition technology weren’t his only remarks to the U.S. Senate about how to guard against the actions of foreign state actors.
When asked, Zuckerberg said he could not guarantee that the Russian government-backed troll farm Internet Research Agency (IRA) had been completely removed from Facebook.
“This is an arms race, and they’re going to keep on getting better at this, and we need to invest in it to keep on getting better at this too,” he said. “As long as there are people in Russia whose job it is to interfere in elections around the world, this is going to be an ongoing conflict.”
Waiting for an AI fix
Throughout five hours of testimony Tuesday, Zuckerberg sprinkled plenty of magic AI pixie dust over his testimony. Hate speech? AI will take care of that. Fake news? AI. Terrorists using Facebook-owned platforms? AI. Content moderation? AI.
Tech companies the world over toss a bit of AI on top of everything these days, but Facebook and its AI research group actually have the chops to make such claims.
To be clear: Zuckerberg has been sprinkling the magic AI pixie dust around for some time now.
Assurances that AI will help cure what ails Facebook have been a part of the social media giant’s narrative since at least last June, when European governments called for more action to address terrorist propaganda and hate speech. Also emphasized in recent months has been that Facebook is bringing in more than 20,000 people to work on security and content, a fact repeated over and over in Congressional testimony.
One problem with AI, however, is that members of Congress are proposing legislation now, while Zuckerberg talked about major returns on AI investments within the next five to 10 years.
If regulation and privacy limitations result in restrictions placed on facial recognition, that could stifle innovation for tech companies in the United States, which already appear to be behind Chinese companies, at a time when companies like Face++ are preparing for expansion outside China.
Exactly how Congress might regulate Facebook and other tech giants is still very much up in the air, but services that master facial recognition can pursue applications beyond surveillance like autonomous driving or biometric payments, or can be embedded in consumer hardware.
If you’ve been following tech news since the 2016 presidential election, you may be among the growing number of people who no longer trust Facebook. It seems like every few weeks for the past six months, there’s another revelation about the number of people who were exposed to fake news or the number of users whose information has been improperly used.
Members of Congress — and the entire world watching Zuckerberg’s testimony — might feel they have little reason to trust what he says. As one member of Congress put it, the trail of apologies from Zuckerberg over the past 14 years seems like a fair indication that self regulation hasn’t worked, but his assessment of computer vision appears reasonable.
If Congress enacts regulation that hampers facial recognition in the United States, companies like Face++ and SenseTime could use their early advantage to dominate this sector of computer vision.