<h2>First Impressions</h2>
If you took a cab in Japan this year, you may have come face-to-face with a tablet that scanned your face in order to learn your gender, age, and other characteristics — and used that data to target you with relevant advertisements.
The facial recognition system made the rounds on Twitter on Saturday after Google privacy engineer Rosa Golijan <a href="https://twitter.com/rosa/status/1119736583521853441">posted a photo</a> of the tablet explaining that it was used in Japanese taxis to &#8220;estimate gender in order to deliver the most optimized content&#8221; — an intriguing preview of how pushy, personalized advertisements could one day roll out elsewhere in the world.
<blockquote class="twitter-tweet" data-width="500" data-dnt="true">
🤨 <a href="https://t.co/uXWOUXz25S">pic.twitter.com/uXWOUXz25S</a>
&mdash; Rosa Golijan (@rosa) <a href="https://twitter.com/rosa/status/1119736583521853441?ref_src=twsrc%5Etfw">April 20, 2019</a></blockquote>
<script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>
<h2>Individual Attention</h2>
Another tweet <a href="https://twitter.com/BenParker140/status/1120326113241325569">posted in response</a> to Golijan shows more details, including how the data is collected and how it&#8217;s used. A QR code in that image links to the website of Japanese e-commerce platform developer DeNA Co Ltd — specifically linking to the <a href="https://dena-taxiad.jp/trust_and_safety/en/?from=tablet">&#8220;Trust &amp; Safety&#8221; page</a> for the company&#8217;s &#8220;Premium Taxi Vision&#8221; service.
Premium Taxi Vision, first deployed in January, is meant to make &#8220;riding a taxi more comfortable and convenient,&#8221; according to a <a href="https://dena-taxiad.jp/news/20190125_tipup_with_kyodo_d_and_rescuenow/">DeNA webpage</a> that we translated using Google. Presumably, that comfort and convenience means saving customers from the sheer embarrassment of watching an advertisement meant for a different gender or age group.
<h2>Opt Out</h2>
DeNA&#8217;s website explains that video of passengers is used to generate a passenger profile — age, presented gender, number of passengers — that gets sent to the advertising agency Geniee, Inc, which then sends a relevant advertisement.
That means Geniee doesn&#8217;t see the footage. In Golijan&#8217;s case, it would have merely known that an adult woman was riding in the cab.
The actual footage, which DeNA claims is immediately deleted, and facial recognition software never leave the tablet.
More on ubiquitous advertisements: <a href="https://futurism.com/pepsi-orbital-billboard-night-sky">Pepsi Plans to Project a Giant Ad in the Night Sky Using Cubesats</a>

The opt-out button is hidden, but footage never leaves the car.

Japanese taxis are using facial recognition to target ads to riders

Now Loading: Girl Ad

Japanese ecommerce firm DeNA has equipped taxis with a facial recognition system that selects ads based on passenger gender and age.

Japanese Taxis Are Using Facial Recognition to Target Ads to Riders

Cameras in the back seat guess your age and gender — and then use them to serve you with advertisements.

<h2>Mixed Messages</h2>
A New York City trial to determine whether <a href="https://futurism.com/amazons-face-rekognition-race-gender" target="_blank" rel="noopener">facial recognition tech</a> could identify drivers as they entered or exited the city via the Robert F. Kennedy (RFK) Bridge appears to be a complete failure that successfully identified zero percent of travelers — but that isn&#8217;t stopping the city from moving forward with the program.
In a <a href="https://www.wsj.com/articles/mtas-initial-foray-into-facial-recognition-at-high-speed-is-a-bust-11554642000?ns=prod/accounts-wsj" target="_blank" rel="noopener">story</a> published Monday, The Wall Street Journal says it reviewed an email sent by a Metropolitan Transportation Authority (MTA) official to a senior official in NY governor Andrew Cuomo&#8217;s administration on Nov. 29.
According to the WSJ, the MTA official wrote that the &#8220;initial period for the proof of concept testing at the RFK for facial recognition has been completed and failed with no faces (0%) being detected within acceptable parameters.&#8221;
The official also wrote, puzzlingly, that the MTA was procuring additional cameras to expand the project to other locations.
<h2>Safety Vs. Privacy</h2>
When questioned about the WSJ&#8217;s story, a Cuomo administration official <a href="https://gizmodo.com/new-yorks-creepy-plan-to-scan-faces-at-toll-booths-has-1833885611" target="_blank" rel="noopener">told Gizmodo</a> that &#8220;the governor’s number one priority is the safety and security of New Yorkers.&#8221;
Daniel Schwarz, privacy and technology strategist at the New York Civil Liberties Union, disagrees.
&#8220;This latest news validates our concerns that this technology is invasive and inaccurate—and the government has no justification for using it to undermine the privacy of our daily commutes,&#8221; he told Gizmodo. &#8220;Facial recognition technology vastly expands the government’s capacity for intrusive, real-time surveillance.&#8221;
READ MORE: <a href="https://www.wsj.com/articles/mtas-initial-foray-into-facial-recognition-at-high-speed-is-a-bust-11554642000?ns=prod/accounts-wsj" target="_blank" rel="noopener">MTA’s Initial Foray Into Facial Recognition at High Speed Is a Bust</a> [The Wall Street Journal]
More on facial recognition: <a href="https://futurism.com/amazons-face-rekognition-race-gender" target="_blank" rel="noopener">Amazon’s Facial Recognition Struggles With Darker Skin</a>

But that isn't stopping the city from expanding the program.

New York’s facial recognition trial was a complete failure

Scan Scam

A facial recognition trial focused on New York's Robert F. Kennedy Bridge didn't correctly identify a single person's face.

New York’s Facial Recognition Trial Was a Complete Failure

The system didn’t correctly identify a single face.

<h2>Eyes Up Here</h2>
The founder of MoviePass is back, and this time he&#8217;s not messing around.
MoviePass co-founder Stacy Spikes has a new startup called PreShow, and it&#8217;s using a business model straight out of &#8220;A Clockwork Orange&#8221;: The company&#8217;s new app helps you score free movie tickets — as long as you sit through 15 to 20 minutes of advertisements that use eye-tracking technology to make sure you don&#8217;t look away.
<h2>Swipe Left</h2>
Unlike YouTube ads, which can be skipped after a few seconds or blocked with a third-party extension, PreShow demands your attention. To use the app, according to <a href="https://www.engadget.com/2019/03/21/preshow-free-movie-tickets/">Engadget</a>, you have to sign in with <a href="https://futurism.com/fbi-unlock-iphone-face">Apple&#8217;s FaceID</a>. And you can&#8217;t just walk away or let your attention wander — the app watches you to make sure you’re paying attention.
To earn the movie tickets, you have to keep your gaze in frame, and look at your screen for the entire duration of the ads. Look away, or put your phone down, and a red border appears, pausing the ads until you give in and return.
<h2>Dystopia-Lite</h2>
Spikes says he doesn&#8217;t plan on licensing the ad-pushing technology anywhere outside of PreShow, Engadget reports.
But that doesn&#8217;t mean executives don&#8217;t plan to cash out. Right now, the app is still gathering backers via a Kickstarter. Once it&#8217;s released, it&#8217;s possible that the app — which, again, forces people to watch a TV-episode length slew of advertisements — could be available only through a monthly subscription.
READ MORE: <a href="https://www.engadget.com/2019/03/21/preshow-free-movie-tickets/">MoviePass founder&#8217;s next big idea: Free movie tickets for watching ads</a> [Engadget]
More on cinema: <a href="https://futurism.com/netflix-future-film-industry">Hollywood is Wrong: Netflix is the Future of Film</a>

What else were you going to do during those 20 minutes?

Brutal startup is using eye tracking to force you to watch ads

All-Seeing App

MoviePass is launching a new app, PreShow, which will force you to watch — actually watch — 20 minutes of ads in exchange for movie tickets.

Brutal Startup Is Using Eye Tracking to Force You to Watch Ads

It's basically like "A Clockwork Orange."

<h2>Move Fast and Break Things</h2>
The countless pictures of people that are used to train facial recognition artificial intelligence systems are often used without consent.
Most notably, IBM downloaded and used almost 1 million pictures from public Flickr accounts without notifying the pictures&#8217; photographers or subjects, <a href="https://www.nbcnews.com/tech/internet/facial-recognition-s-dirty-little-secret-millions-online-photos-scraped-n981921">according to NBC News</a>.
&#8220;This is the dirty little secret of AI training sets,&#8221; NYU law professor Jason Schultz told NBC. &#8220;Researchers often just grab whatever images are available in the wild.&#8221;
<h2>Positive Twist</h2>
When IBM announced that it had created its new dataset, in which each of the pictures was annotated with people&#8217;s measurements and descriptions, <a href="https://www.ibm.com/blogs/research/2019/01/diversity-in-faces/">the company framed it</a> as a way to make facial recognition systems fairer. That&#8217;s because before training AI on user-generated images became common, many algorithms only learned from limited datasets like <a href="https://futurism.com/the-byte/ai-portrait-art">celebrity images taken off of the web</a>.
But the reality isn&#8217;t that rosy. Particularly concerning is how images of people&#8217;s faces are being used without their knowledge or consent to train algorithms that then may be used to surveil and monitor them. For example, <a href="https://theintercept.com/2018/09/06/nypd-surveillance-camera-skin-tone-search/">The Intercept </a><a href="https://theintercept.com/2018/09/06/nypd-surveillance-camera-skin-tone-search/">reported</a> that IBM sold the NYPD surveillance systems that let cops scan CCTV footage and filter by people&#8217;s race.
&#8220;People gave their consent to sharing their photos in a different internet ecosystem,&#8221; Meredith Whittaker, co-director of the <a href="https://ainowinstitute.org/">AI Now Institute</a>, told NBC. &#8220;Now they are being unwillingly or unknowingly cast in the training of systems that could potentially be used in oppressive ways against their communities.&#8221;
<h2>Claiming Dibs</h2>
IBM claims it will accept people&#8217;s requests to take certain images down, but the company doesn&#8217;t make it easy. It only sends the dataset to academic and corporate research teams, so there&#8217;s no way that someone would even know their picture had been used, according to NBC.
NBC reports that IBM made one photographer jump through hoops to get his photos removed, and only deleted four of over 1,000 of his photos that were included in the set.
&#8220;This is the type of mass collection and use of biometric data that can be easily abused,&#8221; class-action lawyer Jay Edelson told NBC, &#8220;and appears to be taking place without the knowledge of those in the photos.&#8221;
READ MORE: <a href="https://www.nbcnews.com/tech/internet/facial-recognition-s-dirty-little-secret-millions-online-photos-scraped-n981921">Facial recognition&#8217;s &#8216;dirty little secret&#8217;: Millions of online photos scraped without consent</a> [NBC News]
More on facial recognition: <a href="https://futurism.com/cops-amazon-facial-recognition">Cops Are Using Amazon’s Facial Recognition Software Wrong</a>

"This is the type of mass collection and use of biometric data that can be easily abused and appears to be taking place without the knowledge of those in the photos."

The face images used to train AI are taken without consent

Finders Keepers

IBM recently published a dataset for facial recognition AI made up of images that it scraped off of Flickr without asking for permission.

The Face Images Used to Train AI Are Taken Without Consent

"This is the dirty little secret of AI training sets."

<h2>Robotic Chivalry</h2>
A newly-published <a href="http://www.freepatentsonline.com/20190039570.pdf">patent</a> shows that Apple obtained the rights to a facial recognition system that would allow your car to recognize you as you approach.
Such a system would let cars unlock their doors when they see a registered user approach — and maybe even remember how each person likes to position the driver&#8217;s seat and mirrors.
<h2>Post-Axing</h2>
Don&#8217;t hold your breath for smart cars that see you coming, though. The patent was published earlier this week, but Apple originally filed it back in early 2017, <a href="https://www.engadget.com/2019/02/12/apple-worked-on-biometric-unlock-for-cars/">according to Engadget</a>.
That was before Apple <a href="https://futurism.com/apple-project-titan-employees-self-driving-car">laid off 200 employees</a> from Project Titan, its self-driving car division, in late January 2019. The company had been working on autonomous vehicles <a href="https://futurism.com/bid-compete-apple-grows-fleet-self-driving-cars">for several years</a> prior to the layoffs, so it&#8217;s unclear whether the facial recognition patent is a relic from Project Titan&#8217;s heyday or if it&#8217;s still part of an active project within Apple&#8217;s research labs.
<h2>Secure-ish</h2>
Facial recognition hooked up to a car door also raises the question of who else gets access to your car. When FaceID made its way to Apple&#8217;s iPhone, it quickly became an easy way for law enforcement agencies to <a href="https://futurism.com/fbi-unlock-iphone-face">bypass passwords</a> — all they had to do was take a person&#8217;s phone and hold it up in front of them.
There&#8217;s no indication if or when this technology will make its way to future cars. But if it does, we&#8217;ll face brand new challenges in the back-and-forth between convenience and privacy.
READ MORE: <a href="https://www.engadget.com/2019/02/12/apple-worked-on-biometric-unlock-for-cars/">Apple worked on biometric unlock for cars</a> [Engadget]
More on facial recognition: <a href="https://futurism.com/amazon-regulation-facial-recognition-tech">Amazon Calls for Regulation of Its Own Facial Recognition Tech</a>

But after massive layoffs in Project Titan, it's unclear whether Apple will develop the tech.

Apple patented facial recognition to unlock your car

Easy Access

A newly-published patent shows that Apple wanted your car to recognize you and unlock the doors when you approach. But it may not pursue the tech.

Apple Patented Facial Recognition to Unlock Your Car

The patent was filed before massive layoffs at Apple’s self-driving car division.

<h2>Turncoat</h2>
Amazon sells one of the most well-known facial recognition services — but now it&#8217;s calling on politicians to develop stronger regulations for the controversial technology.
On Thursday, Amazon Web Services&#8217; Global Public Policy VP Michael Punke <a href="https://aws.amazon.com/blogs/machine-learning/some-thoughts-on-facial-recognition-legislation/">published a blog post</a> that outlined several key areas where policy surrounding facial recognition technology, especially when used by police, could be improved. It&#8217;s a bold public move that means Amazon is joining other prominent tech companies, <a href="https://futurism.com/microsoft-president-face-recognition-tech">especially Microsoft</a>, that have also called for stronger rules about how their own facial recognition systems can be used — suggesting that business leaders are wisening up to how their technology can be abused.
&#8220;Over the past several months, we’ve talked to customers, researchers, academics, policymakers, and others to understand how to best balance the benefits of facial recognition with the potential risks,&#8221; Punke wrote. &#8220;It’s critical that any legislation protect civil rights while also allowing for continued innovation and practical application of the technology.&#8221;
<h2>Punke Rock</h2>
Much of the blog post could be read as an attempt to protect Amazon from mounting criticism over its facial recognition service, Rekognition. For instance, each proposed regulation, such as a call for improved transparency among organizations that use facial recognition tech, is followed by a note about how Amazon is already following its own proposals.
Punke&#8217;s blog seems to insist again and again that Amazon is really the good guy here — the company just wants to make sure that its technology is used responsibly. In the blog, Punke seeks to discredit recent independent research that found that the system <a href="https://futurism.com/the-byte/amazon-rekognition-falsely-matched-congresspeople">could make significant errors</a> and also propagates <a href="https://futurism.com/amazons-face-rekognition-race-gender">racial and gender biases</a>. Punke repeated the company&#8217;s official defense — which <a href="https://futurism.com/cops-amazon-facial-recognition">its own customer has debunked</a> — that the researchers simply used Rekognition wrong.
It&#8217;s this self-serving aspect of Amazon&#8217;s call for regulation that has drawn criticism from the ACLU.
&#8220;Amazon&#8217;s framework rings woefully hollow,&#8221; ACLU Senior Legislative Counsel Neema Singh Guliani <a href="https://www.vox.com/technology/2019/2/7/18216125/amazon-regulation-facial-recognition-software">told Vox</a>. She added that the proposal &#8220;underscores the company&#8217;s refusal to properly address the dangers of its technology in government hands and reinforces the urgent need for Amazon to get out of the surveillance business altogether.&#8221;
<h2>Open Dialogue</h2>
All the same, Punke is correct to assert that facial recognition software and those who use it should be subject to stronger regulations than they are today.
&#8220;New technology should not be banned or condemned because of its potential misuse,&#8221; Punke wrote. &#8220;Instead, there should be open, honest, and earnest dialogue among all parties involved to ensure that the technology is applied appropriately and is continuously enhanced.&#8221;
He has a point; a recent survey found that <a href="https://futurism.com/facial-recognition-survey-americans">Americans are fine with facial recognition software</a> just so long as it works and is used responsibly. All the same, those regulations probably shouldn&#8217;t be drafted by the same company that makes a killing selling facial recognition tools to the police.
READ MORE: <a href="https://www.vox.com/technology/2019/2/7/18216125/amazon-regulation-facial-recognition-software">Amazon is trying to regulate itself over facial recognition software — before Congress does</a> [Vox]
More on Rekognition: <a href="https://futurism.com/the-byte/amazon-facial-recognition-ice">Report: Amazon Tried to Sell Facial Recognition Tech to ICE</a>

But the company also wants you to know that it's the good guy here.

Amazon calls for regulation of its own facial recognition tech

Self-Policing

In a recent blog post, an Amazon VP called for stronger regulations for facial recognition technology, including Amazon's own Rekognition.

Amazon Calls for Regulation of Its Own Facial Recognition Tech

The company may be hoping to pen some of the government's ethical standards. 

<h2>Ban Hammer</h2>
Facial recognition technology is everywhere you look — from <a href="https://futurism.com/apple-just-released-new-details-about-iphone-x-facial-recognition">unlocking phones</a> to <a href="https://www.forbes.com/sites/bernardmarr/2018/12/17/the-amazing-ways-facial-recognition-ais-are-used-in-china/#44411c425fa5">shaming jaywalkers</a>. But should corporations have the power to use it on you without consent?
That&#8217;s the question the city of San Francisco is tackling right now. A member of the city&#8217;s Board of Supervisors proposed a ban on facial recognition technology for city agencies on Tuesday, <a href="https://www.wired.com/story/san-francisco-could-be-first-ban-facial-recognition-tech/">Wired reports</a> — potentially forcing tech companies to justify the use of surveillance tools.
<h2>&#8220;Coercive and Oppressive Ends&#8221;</h2>
San Francisco city board member Aaron Peskin is calling for an approval process for any new surveillance technology purchases by city agencies such as license plate readers, CCTV, and gun-detection systems.
&#8220;I have yet to be persuaded that there is any beneficial use of this technology that outweighs the potential for government actors to use it for coercive and oppressive ends,” Peskin told Wired.
Notably, the proposal doesn&#8217;t include an approval process for the use of private surveillance tech in public spaces. It also leaves questions about how the bill might affect local law enforcement.
<h2>Growing Momentum</h2>
There&#8217;s momentum growing around the idea of limiting the use of facial recognition technology — particularly when it comes to law enforcement. Even Microsoft CEO Satya Nadella warned of a <a href="https://futurism.com/microsoft-president-face-recognition-tech">1984-like future</a> at a Web Summit in November — despite reports of his company <a href="http://fortune.com/2019/01/15/coalition-pressures-amazon-microsoft-google-facial-recognition-surveillance-government/">selling facial recognition tech</a> to the government for surveillance.
But the tech might already be so deeply entrenched in society that it&#8217;s too late to rip it out of the ground and ban it completely.
READ MORE: <a href="https://www.wired.com/story/san-francisco-could-be-first-ban-facial-recognition-tech/">San Francisco Could Be First to Ban Facial Recognition Tech</a> [Wired]
More on facial recognition: <a href="https://futurism.com/the-byte/facial-recognition-red-flags-police-lied">Police Said You Could Skip Public Facial Recognition. They Lied.</a>

The proposal also calls for an approval process for any future surveillance tech.

San Francisco is considering a ban on facial recognition

Face Blindness

A San Francisco Board of Supervisors member proposed to ban facial recognition outright, calling for an approval process for any future surveillance tech.

San Francisco Is Considering a Ban on Facial Recognition

Should the government be forced to justify new surveillance tools?

<h2>Flags Raised</h2>
In December, London&#8217;s Metropolitan Police Service (the Met) <a href="https://futurism.com/the-byte/facial-recognition-scan-deny-suspicious" target="_blank" rel="noopener">announced plans</a> to use facial recognition technology to scan people&#8217;s faces during a public pilot program.
&#8220;Anyone who declines to be scanned during the deployment will not be viewed as suspicious by police officers,” the Met <a href="http://news.met.police.uk/news/facial-recognition-technology-to-be-deployed-on-17-and-18-december-340759" target="_blank" rel="noopener">wrote</a> in its announcement of the program. “There must be additional information available to support such a view.&#8221;
But based on a <a href="https://www.independent.co.uk/news/uk/crime/facial-recognition-cameras-technology-london-trial-met-police-face-cover-man-fined-a8756936.html?utm_source=reddit.com" target="_blank" rel="noopener">story</a> published in The Independent on Friday, that turned out to be a lie.
<h2>Covered Up</h2>
Sources told The Independent they saw police stop several people who had covered their faces during a trial of the tech on January 31.
A group of officers asked one man, identified <a href="https://www.computerweekly.com/news/252456900/Live-facial-recognition-technoloy-has-been-deployed-for-the-ninth-time-by-Metropolitan-Police" target="_blank" rel="noopener">by Computer Weekly</a> as Chris Johnson, to hand over his identification, which he did, according to Silkie Carlo, director of a nonprofit focused on privacy protection, who was on site during the trial.
The police them became &#8220;accusatory and aggressive,&#8221; Carlo told The Independent. &#8220;The guy told them to piss off and then they gave him the £90 public order fine for swearing.&#8221;
&#8220;I don’t want my face on anything — I’m not a criminal,&#8221; Johnson told Computer Weekly. &#8220;If I want to cover my face, I’ll cover my face, but the police officer, because I swore, gave me a fine — all because I decided to walk down the street and I didn’t want them looking at my face, which I think I’ve got a right to do.”
<h2>Cover Up?</h2>
Carlo told The Independent Johnson didn&#8217;t do anything to warrant the stop.
&#8220;He simply pulled up the top of his jumper over the bottom of his face, put his head down, and walked past,&#8221; she said. &#8220;There was nothing suspicious about him at all.&#8221;
The Met had a different take on the encounter.
&#8220;Officers stopped a man who was seen acting suspiciously in Romford town centre during the deployment of the live facial recognition technology,&#8221; the organization wrote in a statement, according to The Independent. &#8220;After being stopped, the man became aggressive and made threats towards officers. He was issued with a penalty notice for disorder as a result.&#8221;
Unless the Met is willing to elaborate on how Johnson was &#8220;acting suspiciously&#8221; — or better yet release video of this suspicious behavior — it certainly sounds like officers stopped him because he was attempting to avoid having his face scanned. Which is exactly the sort of thing the Met promised it wouldn&#8217;t do.
READ MORE: <a href="https://www.independent.co.uk/news/uk/crime/facial-recognition-cameras-technology-london-trial-met-police-face-cover-man-fined-a8756936.html?utm_source=reddit.com" target="_blank" rel="noopener">Police Stop People for Covering Their Faces From Facial Recognition Camera Then Fine Man £90 After He Protested</a> [The Independent]
More on facial recognition: <a href="https://futurism.com/the-byte/facial-recognition-scan-deny-suspicious" target="_blank" rel="noopener">Denying Facial Recognition Scan Won’t Raise Red Flags, Say Police</a>

The police stopped people who covered their faces, according to reports.

Police said denying facial recognition scans wouldn’t raise red flags. They lied.

Broken Promise

During a facial recognition trial, London police stopped a man for "suspicious behavior," but the man claims all he did was cover his face.

Police Said You Could Skip Public Facial Recognition. They Lied.

A London trial did not go the way police promised it would.

<h2>Riddled With Holes</h2>
When Amazon&#8217;s facial recognition system Rekognition came under fire by studies that found <a href="https://futurism.com/the-byte/amazon-rekognition-falsely-matched-congresspeople">it was inaccurate</a> and biased against <a href="https://futurism.com/amazons-face-rekognition-race-gender">women and people of color</a>, the <a href="https://aws.amazon.com/blogs/machine-learning/thoughts-on-recent-research-paper-and-associated-article-on-amazon-rekognition/">company responded</a> by saying that the researchers misrepresented how Rekognition <a href="https://docs.aws.amazon.com/rekognition/latest/dg/considerations-public-safety-use-cases.html">might be used</a> by Amazon&#8217;s law enforcement clients.
But now, a law enforcement client has come forward and said that no, it uses the software much like the researchers did. Oregon&#8217;s Washington County Sheriff&#8217;s Office, the only enforcement office that Amazon has officially acknowledged to be a client, says that it does not use Rekognition the way that Amazon recently claimed to have recommended, <a href="https://gizmodo.com/defense-of-amazons-face-recognition-tool-undermined-by-1832238149">according to Gizmodo</a>. Needless to say, that&#8217;s worrisome.
<h2>Voluntary Safety Measures</h2>
Amazon claimed that it recommends police set a strict minimum confidence threshold for matches, meaning someone&#8217;s face would have to be at least 99 percent the same as a database image in order to trigger a match.
But the Oregon police said that they don&#8217;t bother using such thresholds, a clarification that&#8217;s backed up by <a href="https://futurism.com/police-surveillance-gets-terrifying-boost-from-amazons-ai-facial-recognition">previous documentation showing a 96 percent match</a>. The Washington County Sheriff&#8217;s Office also clarified that it only uses Rekognition to generate leads and that detectives are required to gather additional evidence before making a case against someone, according to Gizmodo.
Amazon responded to Gizmodo by <a href="https://fruitsoftheweb.kinja.com/amazon-statement-on-rekognition-1832245007">once more asserting</a> that the company collaborated with the Washington County Sheriff&#8217;s Office on deploying Rekognition and that thus far, there had been no public complaints about the tool.
<h2>Unleashed</h2>
What the Washington County Sheriff&#8217;s Office&#8217;s admission means is that at least one client — the only one that&#8217;s publicly known — isn&#8217;t following Amazon&#8217;s guidelines for the software.
And that&#8217;s a worrisome glimpse of the ways police are going to use facial recognition software going forward.
READ MORE: <a href="https://gizmodo.com/defense-of-amazons-face-recognition-tool-undermined-by-1832238149">Defense of Amazon&#8217;s Face Recognition Tool Undermined by Its Only Known Police Client</a> [Gizmodo]
More on Rekognition: <a href="https://futurism.com/the-byte/amazon-rekognition-falsely-matched-congresspeople">Amazon Rekognition Falsely Matched 28 Members of Congress to Mugshots</a>

Cops are using Amazon's facial recognition software wrong

Instruction Manual Not Included

Amazon responded to recent criticism its facial recognition system, Rekognition, by claiming safeguards were in place. The cops using it disagree.

Cops Are Using Amazon’s Facial Recognition Software Wrong

The revelation pokes holes in Amazon's defense of Rekognition.

<h2>We&#8217;ve Been Over This</h2>
Amazon apparently still hasn&#8217;t fixed the problems with its facial recognition software, Rekognition.
The software, which is used by police and <a href="https://futurism.com/the-byte/amazon-facial-recognition-ice">Immigration and Customs Enforcement</a>, has drawn criticism for frequent errors, like when it matched Congresspeople&#8217;s headshots with <a href="https://futurism.com/the-byte/amazon-rekognition-falsely-matched-congresspeople">photos from a mugshot database</a>. Now, new research shows that Rekognition generally fares well with white men&#8217;s faces, but it struggles to identify light-skinned women and anyone with darker skin, <a href="https://www.theverge.com/2019/1/25/18197137/amazon-rekognition-facial-recognition-bias-race-gender">according to The Verge</a>.
<h2>Gender Bias</h2>
Presented with the faces of light-skinned women, Rekognition incorrectly labeled 19 percent of them as men, <a href="http://www.aies-conference.com/wp-content/uploads/2019/01/AIES-19_paper_223.pdf">according to new research</a> from MIT and the University of Toronto. Women with dark skin were incorrectly labeled as men 31 percent of the time, according to the research, which will be presented this weekend at the <a href="http://www.aies-conference.com/">AAAI/ACM Conference on Artificial Intelligence, Ethics, and Society</a>.
Amazon responded to the research similarly to how it dismissed past criticisms of Rekognition — the company argued that the researchers hadn&#8217;t used the latest version of Rekognition, according to The Verge. Amazon also pushed back on the basis that the facial analysis system that detects facial expressions and gender is separate from the facial recognition system that police might use to compare people to a mugshot. Fair points — but not ones that address the actual problem detected by the researchers head-on.
<h2>Broken System</h2>
Ultimately, the researchers hope to accomplish more than encouraging surveillance systems that can accurately detect people of all races and genders — they also want to bring about a society in which such technology is properly regulated to prevent police or <a href="https://futurism.com/amazon-rekognition-employees">anyone else from abusing people</a>.
&#8220;Furthermore, while algorithmic fairness may be approximated through reductions in subgroup error rates or other performance metrics, algorithmic justice necessitates a transformation in the development, deployment, oversight, and regulation of facial analysis technology,&#8221; the researchers wrote in their paper.
READ MORE: <a href="https://www.theverge.com/2019/1/25/18197137/amazon-rekognition-facial-recognition-bias-race-gender">Gender and racial bias found in Amazon’s facial recognition technology (again)</a> [The Verge]
More on Rekognition: <a href="https://futurism.com/police-surveillance-gets-terrifying-boost-from-amazons-ai-facial-recognition">Police Surveillance Is Getting a Helping Hand from… Amazon!</a>

The company's AI still seems particularly biased.

Amazon's facial recognition struggles with darker skin

Racist Robots

Amazon's facial recognition software, Rekognition, struggles to identify the gender of women with darker skin, suggesting ongoing algorithmic bias.

/Facial Recognition