Name: Insights: Microsoft Wants Regulation for Facial Recognition Tech
Uploaded: 2020-04-14
Duration: 3 min 56 s
Description: The company's chief legal officer explains why regulation would be good for this type of tech.

 BRAD SMITH: I think there's three things quickly worth noting about facial recognition. First, I think it's becoming sort of the first specific concrete application of artificial intelligence, where you're seeing real controversy. And therefore, it's becoming a test bed for how governments will think about regulating it. Second, we ourselves at Microsoft proposed regulation of this technology last July, just 11 months ago. We said, because of the risks of bias, issues around privacy, democratic freedoms. This needs to be regulated by the law. And 11 months ago when we did that, we got feedback and blowback in Silicon Valley. What's wrong with you guys? You must be behind. Why do you want the government involved in this? And then, within less than a year, the city next to Silicon Valley enacted this new law. And what the law does, is it imposes a moratorium on all public sector use in San Francisco by the city of facial recognition technology. I think that's a bit of a challenge in its own right, because ultimately there are beneficial uses of this technology. It is being used to help diagnose certain diseases. It is being used to reunite people who have gone missing with their families. You know, it will be used in appropriate and important ways to keep the public safe. So a moratorium is one thing, but ultimately you've got to make decisions. Cities, states, countries are going to have to make decisions. How does it want this technology to be used? ADAM LASHINSKY: And so I envision that Microsoft will be part of a group that will go to jurisdictions like the city of San Francisco, and say, let's talk about this, and figure out how you can implement this technology. How long is that conversation going to take do you think, while under-- I think you've sort of alluded to the fact that the Chinese and others are implementing, implementing, implementing, right? BRAD SMITH: Yeah, well, of course right now this technology except for the city of San Francisco is unregulated in the United States as well. It's mostly unregulated in the world, although, certainly in Europe, it is subject to the general privacy laws that are there. I think it's something that needs to move forward faster, rather than more slowly. We've proposed legislation. We took a draft of legislation to the Washington. Senate in the house. It got through the Senate, but died in the house. We are where-- I'm basically traveling the world among other things encouraging governments to start legislating. And our basic proposition is-- look, you can't answer yet every question that the world is going to need to answer in the future. But there are some questions that we believe can be answered clearly now. Give you one example. Right now, there is no legal requirement that somebody who enters the facial recognition business needs to make their service available for testing. So you don't have a Consumer Reports. You don't have academics that can test these services for bias, and yet there's documented instances of facial recognition being biased against women and people of color. That's the kind of bias that can result in adverse decisions being made that impact people's lives. So one of our proposals has been-- look, if you want to participate in this market, you shouldn't have to make your service available for third party testing in an objective way. That's what automakers do. That's what drug makers do. It's widespread throughout our society. And I'm hopeful that we'll get at least one major jurisdiction somewhere to pass that. All it takes is one, because if it's available for testing somewhere, the results will be readable everywhere. 

Insights: Microsoft Wants Regulation for Facial Recognition Tech

Uncategorized

The chief legal officer says regulation would be good tech.

Microsoft Wants Regulation for Facial Recognition Tech

The company's chief legal officer explains why regulation would be good for this type of tech.

I think there's three things quickly worth noting about facial recognition. First, I think it's becoming sort of the first specific concrete application of artificial intelligence where you're seeing real controversy on. Therefore, it's becoming a test bed for how governments will think about regulating it. Second, we ourselves at Microsoft proposed regulation of this technology last July. Just 11 months ago, we said, because of the risks of bias issues around privacy, democratic freedoms, this needs to be regulated by the law. And 11 months ago, when we did that, we got feedback and sort of blowback in Silicon Valley. What's wrong with you guys? You must be behind. Why do you want the government involved in this? And then within less than a year, the city next to Silicon Valley enacted this new law, and what the law does is it imposes a moratorium on all public sector use in San Francisco by the city of facial recognition technology. I think that's a bit of a challenge in its own right because ultimately there are beneficial uses of this technology is being used to help diagnose certain diseases. It is being used to reunite people who have gone missing with their families. You know, it will be used inappropriate and important ways to keep the public safe. So a moratorium is one thing. But ultimately, you've got to make decisions. Cities, states, countries. We're gonna have to make decisions. How does it want this technology to be used? And so I envision the Microsoft will be part of a group that will go to jurisdictions like the City of San Francisco and say, Let's talk about this and figure out how you can implement this technology. How long is that conversation gonna take? Do you think while under you I think you sort of alluded to the fact that the Chinese and others are implementing, implementing, implementing, right? Yeah, well, and of course, right now, this technology, except for the city of San Francisco, is unregulated in the United States as well, mostly unregulated in the world. Although certainly in Europe it is subject to the general privacy laws that are there. Um, I think it's something that needs to move forward faster rather than more slowly. We proposed legislation. We took a draft of legislation to the Washington, uh, Senate and house you got through the Senate but died in the house. We will work. I'm basically traveling the world, among other things, encouraging governments to start legislating. And our basic proposition is, Look, you can't answer yet every question that the world is going to need to answer in the future. But there are some questions that we believe can be answered clearly. Now give you one example. Right now, there's no legal requirement that somebody who enters the facial recognition business needs to make their service available for testing. So you don't have a consumer reports. You don't have academics that contest the service is for bias. And yet there's documented instances off facial recognition being biased against women and people of color. That's the kind of bias that can result in adverse decisions being made that impact people's lives. So one of our proposals have spent Look, if you want to participate in this market, you should have to make your service available for third party testing in an objective way. That's what automakers do. That's what drug makers do. It's widespread throughout our society, and I'm hopeful that we'll get at least one major jurisdiction somewhere to pass that all it takes is one, because if it's available for testing somewhere, the results will be readable everywhere