Facial Recognition

Microsoft and Amazon Adopt Temporary Bans on Police Use of Their Facial Recognition Tech. That's Not Nearly Enough.

Both companies need to join IBM and others in entirely abandoning the development of this mass surveillance technology.


Microsoft CEO Brad Smith declared during an online Washington Post event today that his company "will not sell facial-recognition technology to police departments in the United States until we have a national law in place, grounded in human rights that will govern this technology." This follows Amazon's announcement yesterday that it is implementing a one-year moratorium on police use of the company's controversial Rekognition facial recognition platform. Amazon says it hopes Congress will step up during this pause to implement appropriate rules for facial recognition technology.

These moves follow IBM CEO Arvind Krishna's letter to Congress announcing that his company is getting out the facial recognition business. "IBM firmly opposes and will not condone uses of any technology, including facial recognition technology offered by other vendors, for mass surveillance, racial profiling, violations of basic human rights and freedoms, or any purpose which is not consistent with our values," Krishna writes. "We believe now is the time to begin a national dialogue on whether and how facial recognition technology should be employed by domestic law enforcement agencies."

The digital rights group Fight for the Future has issued a statement calling Amazon and Microsoft's moves "essentially a public relations stunt." The companies' researchers will "spend the next year 'improving' the accuracy of their facial recognition algorithms, making it even more effective as an Orwellian surveillance tool," the group warned. "The reality is that facial recognition technology is too dangerous to be used at all….Congress should act immediately to ban facial recognition for all surveillance purposes."

While Microsoft and Amazon express their pious hopes that Congress might soon pass comprehensive legislation to "implement appropriate rules" for deploying and using facial recognition technology, a quick scan of the relevant introduced bills finds little pending action on the issue.

Although Congress continues to dawdle over facial recognition legislation, the California legislature is considering a bill that a coalition of civil liberties organizations led by the American Civil Liberties Union (ACLU)  warns would "legitimize the widespread use of harmful and unnecessary facial recognition on the public."

In a letter to the legislature, the civil liberties groups argue that the California bill "allows governments to identify, locate, and track people using facial recognition, a technology that gives governments the unprecedented power to spy on us wherever we are—identifying us at protests, doctor's appointments, political rallies, places of worship, and more." Moreover, the legislation would undercut the outright bans on police use of facial recognition that have been adopted by several California cities, including San Francisco, Oakland, and Berkeley.

Meanwhile the secretive surveillance company Clearview AI is taking an entirely different tack—it is jettisoning all of its private business clients and will sell its facial recognition services only to law enforcement agencies. Clearview AI has created an app that enables police to match photographs to its database of over 3 billion photos scraped from millions of public websites including Facebook, YouTube, Twitter, Instagram, and Venmo.

In a 2019 FAQ no longer available to the public through the Clearview AI site, the company claimed that it

has the most accurate facial identification software in the world, with a 98.6% accuracy rate. This does not mean that you will get matches for 98.6% your searches, but you will almost never get a false positive. You will either get a correct match or no results. We have a 30-60% hit rate, but we are adding hundreds of millions of new faces every month and expect to get to 80% by the end of 2019.

Earlier this month the ACLU filed a lawsuit in Illinois alleging that Clearview AI has violated the state's Biometric Information Privacy Act that forbids companies from using a resident's face scans without their consent. In its court filings, the ACLU says that Clearview AI's business model "appears to embody the nightmare scenario" of "a private company capturing untold quantities of biometric data for purposes of surveillance and tracking without notice to the  individuals affected, much less their consent."

As demonstrations over the police slaying of George Floyd erupted across the country, Sen. Edward Markey (D–Mass.) sent a letter to Clearview AI asking if the company's facial recognition app is currently being used by police to identify protestors.

"As demonstrators across the country exercise their First Amendment rights by protesting racial injustice, it is important that law enforcement does not use technological tools to stifle free speech or endanger the public," wrote Markey. "The prospect of such omnipresent surveillance also runs the risk of deterring Americans from speaking out against injustice for fear of being permanently included law enforcement databases."

"Facial recognition is the perfect tool for oppression," warn Woodrow Hartzog, a professor of law and computer science at Northeastern University, and Evan Selinger, a philosopher at the Rochester Institute of Technology. The technology, they add, is "the most uniquely dangerous surveillance mechanism ever invented."

These temporary moratoria by Microsoft and Amazon are wholly inadequate responses to the dangers this technology poses to civil liberties. Both companies should join IBM in entirely abandoning a technology that the national security state is sure to abuse.

NEXT: Will the Cops Who Killed Kenneth Chamberlain After Illegally Breaking Into His Apartment Ever Be Held Accountable?

Editor's Note: We invite comments and request that they be civil and on-topic. We do not moderate or assume any responsibility for comments, which are owned by the readers who post them. Comments do not represent the views of Reason.com or Reason Foundation. We reserve the right to delete any comment for any reason at any time. Report abuses.

  1. While we are at it, how about we talk about Google’s work enabling the Chinese police state to control their people’s access to the internet. If we are going to start going after Tech companies for doing work we don’t like for the police in the US, then we sure as hell also talk about prohibiting them doing work for foreign police states like China.

    1. Yep; my first thought was “what are they going to do about all the rest of the world?”
      It would be fun to watch the panic if the SEC required all companies to disclose what percentage of their gross and net profits came from Communist China. Splitting out finished goods and components.

      1. ” Online Jobs 2020 Here ”

        I’ve made $66,000 so far this year w0rking 0nline and I’m a full time student. I’m using an 0nline business opportunity I heard about and I’ve made such great m0ney.PDs It’s really user friendly and I’m just so happy that I found out about it.

        This is what I do……… More Detail

    2. While we are at it, how about we talk about Google’s work enabling the Chinese police state to control their people’s access to the internet.

      Nah, that’s fine bro. Talking about that might impact FREE TRADE.

      1. I am now making extra $19k or more every month from home by doing very simple and easy job online from home. I have received exactly $20845 last month from this home job.xcx Join now this job and start making extra cash online by follow instruction

        on the given website………… See More here

    3. John, my attitude is: Well boys, it is a little late to develop a conscience. Too late, that ship sailed.

      The question becomes do we want to toss facial recognition technology (and future applications of that technology) while our adversaries (China, Russia, Iran) have it, use it, and improve it?

      I submit that the answer is No.

      1. Bullshit. What use can facial recognition technology have other than for mass surveillance? Because the bad guys live in a panopticon, we have to also?

        1. Lilly Wood am now getting paid every month more than $31,000 by doing very easy job online from home. I have earned last month $31540 from this easy job just by giving this job only 2 to 3 hrs a day using my laptop. Everybody on this earth can now get this job and start making more cash online just by follow instructions on this web page… More Reading

  2. Need to send the address of Clearview AI over to ANTIFA…

    1. So they can beg them for food to support their autonomous zone?

  3. How about Microsoft, Amazon and every other tech company show some real balls and say “we don’t care what laws are in place, we don’t believe there’s any government that can use such technology responsibly and we’re not going to develop or sell anything to them since it will inevitably be used for evil.”

    1. They are run by statists; they just want to be sure their products are not used by anyone right of Stalin. If they really gave a damn, they would not develop the technology.

    2. Possible government response: “Actually, we don’t need your half-assed technology now that we’ve mandated wearing license-plated facemasks.”

      1. I’m well aware that the government is going to get it anyway. I’m just sick of seeing all of these corporate symbolic half-measures. Giving out millions of dollars of donations to BLM and other groups while simultaneously pretending not to feed the states ability to rule over people is nonsense.

        Speaking of these mega donations, I can’t wait to read Reason’s investigation about how that money is all being distributed once it reaches these groups. I have absolutely no doubt that every penny will be used to help all people of color equally. No corruption here!

        1. I’m well aware that the government is going to get it anyway. I’m just sick of seeing all of these corporate symbolic half-measures.

          These corporations are creations of the state; their meteoric rise in stock price is a result of governmental policies; their near monopolies are the result of government regulations; they are in bed with one of the major parties. All their decisions are driven by making those with political power happy.

          Speaking of these mega donations, I can’t wait to read Reason’s investigation about how that money is all being distributed once it reaches these groups. I have absolutely no doubt that every penny will be used to help all people of color equally. No corruption here!

          BLM is explicitly a political organization, not a charity that helps “people of color” on the ground. Why would you expect them to distribute any funds?

          1. I’m 100% with you on the corporations being symbiotic creations of the state.

            I don’t expect BLM to distribute anything, that’s my point. Such groups are receiving what I imagine now amounts to hundreds of millions of dollars based on some of the donations ($10M from Amazon alone).

            Here’s a hilarious article:

            Look at these good corporate sponsors. Don’t let the headline fool you, it doesn’t tell you anything about where the money is actually going, it’s a consolidated list of bragging rights for corporations. It’s an ad to say who the good guys are. Ok, the NAACP just got tens on millions of dollars. Now what?

            Even more hilarious is that they felt the need to add this little blurb about Facebook:

            “However, this response has largely been overshadowed by internal division over Zuckerberg’s decision not to flag or remove a number of President Donald Trump’s posts.”

            Was it overshadowed because NBC news said it was? What’s the proper ransom to look like a good guy Mr. Peacock?

            1. I don’t expect BLM to distribute anything, that’s my point.

              And my point is that that isn’t “corruption”; BLM is explicitly a political advocacy organization, and we should treat them as such. Corporations that donate to them should be associated with the (neo-Marxist) political agenda BLM promotes.

      2. I think people will veil their faces in the future. Physical veils or electronic veils meant to fool cameras; the thirst for privacy may return someday.

        1. Yeah. My first thought was a snarky comment about masks, then I thought “somebody will make millions off a necklace that blocks your face from cameras”, then my next thought was “damn, I wish I knew how to invent one”.

    3. You’re still stuck in the thinking that big companies control technology; none of those companies do.

      And just because you consider something to be “evil” doesn’t mean everybody in the entire world does.

      1. I’m just fine with moral relativism. Let’s hear an argument for why the state knowing where everyone is at all times and extensively documenting the movements of all citizens is a big win that won’t be susceptible egregious abuses by secretive agencies of the state. Even without the near 100% chance that it gets abused, tell me how it’s a win for the citizenry. Do we get to do it to them? I didn’t think so.

        1. Is this goalpost moving? I honestly can’t tell. But if anyone is going to develop powerful tech, I’d prefer it be the pantywaists in SV than the autocratic thugs in China.

          1. I think NOYB2 is right, SV probably doesn’t have much control over their own tech and how it’s used. I’m not sure it matters who develops it. Neither China nor the US government should have it. Obviously should has nothing to do with it.

            I wasn’t goal post moving, he argued the evil is a subjective term. Ultimately all morality is. Fair enough, but at least give me an argument for why this isn’t a good use of the term if you’re going to bother pointing that out.

            1. I’m not a relativist. But thank you for addressing his first point; that was what irked me.

              Google (God help us) will probably not be responsible for general AI (which is what we’re ultimately talking about; these are machine learning ventures). There are myriads efforts, many of them small, some of them invisible and secret, to conjure that demon. Throwing a lot of money at it helps, but in the end it is individuals that are innovating, not companies. If Silicon Valley falls into a sinkhole tomorrow, surveillance and AI development will continue.

            2. I wasn’t goal post moving, he argued the evil is a subjective term.

              I didn’t do that at all. I just pointed out that you are assuming something (“X is evil”) without evidence or argument to support your view.

  4. They’ll change their tune about selling if there is a Biden administration.

  5. Both companies need to join IBM and others in entirely abandoning the development of this mass surveillance technology.

    Not the take I expected from Bailey. Why should they abandon it? It is going to be developed one way or another. I agree legislation limiting or curtailing it’s use is badly needed. Better it be developed “in the open” and available to consumers as opposed to the Clearview AI model.

    In the end, this is the future. It’s a future of less privacy in meat space. People bring cameras and microphones into their homes all the time. No one can prevent their likeness being captured, without immuring themselves. Hopefully things like neural nets generating fake photos of people will mix things up and make it harder for these recognition programs to identify people.

    1. things like neural nets generating fake photos of people

      Examples for the uninitiated.

      1. Yeah, the face of hunger in America looks (un)surprisingly well fed.

        1. Surprisingly, um, non-minority also.

        2. And remarkably like a gender neutral Justin Trudeau.

      2. Everyday I learn something new. And then shit my pants.

        1. Well, hold on to your butt. This is just baby tricks; these processes (generative adversarial network or GAN) are fairly new. The infant is just teething right now.

      3. That one looks real; the face of hunger is still in the uncanny valley.

    2. “Hopefully things like neural nets generating fake photos of people”

      Holy cow! Business opportunity!

    3. It is an example of Mr. Bailey not using his noggin, and thinking the topic through. I thought he was the science writer. The problem is that his writing shows that he has a tenuous connection to reason and logic.

  6. Both companies need to join IBM and others in entirely abandoning the development of this mass surveillance technology.

    That horse left the barn about five years ago. Face recognition is mature and widely available. Any improvements to it will just happen automatically as part of general AI improvements whether any big company invests in it or not.

    The only limits you can realistically put on face recognition is on its use in government. You can prohibit police from using it and audit that (if you think that’s a good idea). You can and should prohibit its use as evidence in court (since people can just look at the people involved).

    It’s fascinating that a libertarian who pretends to stand up for the 2A thinks he can get away with demanding limiting the development of face recognition. Ron, your mask slipped a long time ago.

    1. I wish Bailey would get in here and explain his position. This is the guy that paid to his genome to be sequenced, right? Why would he be cavalier about genomic data mining but alarmist about photographic data mining?

      1. Because he’s a stupid piece of shit.

  7. Doesn’t matter. The police will just turn to Chinese companies to provide it.

    1. Exactly. I think we need a constitutional amendment. It’s been a while since we’ve passed one, and this seems like something most sides would agree on.

  8. There is, I think, a profound misunderstanding among criminal justice purveyors and their supporters about what, exactly, people are so angry about. People are not angry at most police officers personally – to be so would be foolish and myopic. We are, instead, angry at the system – the bureaucracy – that exists, in large part, to do naught but ensure its own existence and ever-creeping growth at the expense of those they supposedly serve. The system exists to perpetuate the system and that perpetuation comes at the cost human life, yes, but also in the form of exorbitant fees, fines and levies for low-level and victimless “crimes” that are committed daily by a populace who could never hope to know or understand the countless laws, rules, restrictions and regulations that have been foisted upon them over the last 231 years. People are angry that the poorest and most marginalized among them are the ones who bear the burden of the system. They are angry about a system that promises justice but delivers the opposite. They are angry that the agents of that system, be they judges, lawyers, prosecutors or police, are much more concerned about the system itself than they are the justice that is supposed to come before it.

    1. Unfortunately that is a lost cause see Pournelle’s iron law of bureaucracy

      1. “The bureaucracy is expanding to meet the needs of the expanding bureaucracy.”

        1. The first job of the system is to protect the system.

          1. Upon completion of the expansion, we can begin the next phase of expansion as a bridge to following expansion.

    2. I wish I could say that this is all self-apparent and entirely predictable to anyone paying even the least bit of attention. Unfortunately, people believe in the platitudes of stump speeches over the gross reality of a system that oscillates between complete indifference and unrepentant malice.

    3. People are angry about all sorts of different things, and are jumping in on the mass temper tantrum fad

  9. It’s too bad the energy spent vacilating between facial recognition and deep fakes can’t be harnessed more productively.

  10. It’s getting creepy how fast Corporations are falling in line with the televised Revolution.

    1. Of course they are. They don’t care who’s in charge or who they have to make propaganda for as long as they stay on the top of the food chain.

      1. Because this is the first revolution against the people.
        Look at the trend of the last 5 years – populism, more local sovereignty, rejection of “experts” and technocrats, anti centralization…
        And now?
        Now the system throws a temper tantrum so the system can reassert control.
        It’s all bullshit.
        All these anti police, anti racism want bigger government, more rules, more globalism, more racism and segregation.
        The corporations love sheep

        1. This is precisely what I’ve been arguing. CNN, NYT, et all… aren’t covering from the protesters because they’re helping the revolution. They’re feigning sympathy in hopes that it will further marry the angry populace to their team and focus more of their hatred at one another.

          The team leaders on each side suggest the answer lies in making the club bigger so they can take a swing at the other side. Both sides of the angry populace agree and hand more money, power and faith over to their side. Only they’re not really different sides. They’re all one big corporate statist authoritarian wrecking ball.

          There’s a reason why they fight over “anti-lynching bills” when the cameras are rolling but the patriot act gets reupped and expanded unanimously without debate every single time. There’s a reason why, despite all the hemming and hawing when Trump ascended to the thrown, the government didn’t collapse. Practically speaking there’s not much of a difference in day to day life. The “democracy” we have is facile. We’re only aloud to the debate table for the things that don’t really matter and everything else is decided behind closed doors. The leviathan grows no matter who’s feeding it.

  11. Is this the same facial recognition software that gives false positives on black faces? From what I’ve read, I haven’t seen any facial recognition programs that do anywhere close to as well with minority faces. They seem to only work really well on white people for some reason.

    1. I’ve already heard someone argue that facial recognition is racist for that reason. I will happily let them have that privilege exclusively if they really want it. Not sure why anyone would. I am sure that it doesn’t matter what any wants, it’s happening.

      1. It’s not racist, a computer can’t be racist. One might make the argument the programmers were somehow racist, but more likely the parameters you need to make facial recognition work simply aren’t very predictive on certain groups because of fewer facial markers.

        I mean, yeah it’s considered racist to say ‘I have a hard time telling black people apart’ but maybe that stereotype is closer to reality than some might care to admit. Or maybe the programmers in silicon valley really are racists. Or maybe they’re just lazy. I can’t pretend to know.

        No matter the reason or cause for the inaccuracy, one can be sure that if facial recognition ends up being used in court there are going to be a lot of black people wrongly accused as things currently stand.

        Feature, or bug?

        1. “Whites” are mutts.
          The product of the greatest amount of genetic mixing on the planet.
          Primarily, the was determined geographically – Europe is the end of the line for migration out of Africa and Asia, where the land bottles up.
          Western history is the long story of continuous movement, invasion, conflict, and mixing.
          Logically, that leads to “white” people having the greatest variety of traits, which can then be used to create unique points for identification.

          1. It’s really just eye and hair color. None of the other “races” or groups of anatomically similar humans has as much variety in coloration. I have brown eyes, my brother has blue eyes.

          2. Actually, the greatest genetic variation is found in Africa. Europeans and Asians all share a small set of common ancestors.

            1. Much of the genetic variation in Africa is found in small, isolated populations. Meaning, Africans can statistically be less variable than Europeans, even though more genetic variants are found on the African continent.

    2. They seem to only work really well on white people for some reason.

      It’s really no great mystery. White faces make facial features stand out more, improving face recognition and interpretation of facial expressions by both humans and machines.

      There are secondary effects, like familiarity and size of training set, but the advantage for white faces generally remains even when you eliminate those.

  12. Yeah, IBM sure suspended their program alright. Also there has never been any mass data collection on American citizens, I’ll only put in the tip, I won’t cum in your mouth, and the check is in the mail.

  13. Meh.

    Once the corona virus mask thing becomes permanent, just add a hat and sunglasses and reading your face become a chimera.

    On the other hand, we will have to log in and out of every place we go for pandemic tracking, so its a wash.

  14. Abandoning dangerous technology never works, because someone even more dangerous won’t abandon it.

  15. I am making a good MONEY (500$ to 700$ / hr )online on my Ipad .Last month my pay check of nearly 30 k$.This online work is like draw straight-arrow and earn money.Do not go to office.I do not claim to be others,I just work.You will call yourself after doing this JOB,It’s a REAL job.Will be very lucky to refer to this WEBSITE.I
    hope,you can find something,Simply go to the below SITE.GOOD LUCK COPY THIS WEBSITE……….. GoodProfits.Com

  16. There will always be some tech company willing/wanting to sell this tech to the government. So, how ’bout we ban the governments use of it?

  17. Why are Microsoft and Amazon developing Facial Recognition Tech to begin with? Is it to identify what you purchase and sell the info outside companies to bombard you with “buy me” ads? Why is there Siri, Alexa or Cortana to spy on everything you do? Why are there smart appliances? I feel increasingly like a tagged wildebeest having every move I make traced and not just by the government.

    I read there is a company, possibly Google, right now employing scientists to translate brainwaves to language. Wait until they start offering implanted chips that will allow www access without the use of any hardware. Wait until they get inside your head.

  18. What everyone fails to understand is that this technology is not difficult to develop. The vast majority of research is on improving accuracy.

    The government may want to indulge in face recognition and it will only be the will of the people that will prevent it. And there is a fair chance it will happen.

    But that’s not where it will be used. Everyone will eventually have AR glasses and it will simply be too beneficial not to be able to see the stats and metadata on every person you walk by.

    That technology won’t even be developed by large corporation in fear that of the bad PR. No, instead it’s going to be developed by some person who decided that they just wanted to create their generation’s facebook. And it will be everywhere.

Please to post comments

Comments are closed.