{"id":2312,"date":"2021-01-14T20:41:19","date_gmt":"2021-01-14T20:41:19","guid":{"rendered":"https:\/\/thenextweb.com\/?p=1334215"},"modified":"2021-01-14T20:41:19","modified_gmt":"2021-01-14T20:41:19","slug":"stanford-team-behind-bs-gaydar-ai-says-facial-recognition-can-expose-political-orientation","status":"publish","type":"post","link":"https:\/\/www.londonchiropracter.com\/?p=2312","title":{"rendered":"Stanford team behind BS gaydar AI says facial recognition can expose political orientation"},"content":{"rendered":"\n
<\/div>\n

Stanford researcher Michael Kosinski, the PhD behind the infamous \u201cGaydar\u201d AI<\/a>, is back with another phrenology-adjacent (his team swears it\u2019s not phrenology) bit of pseudo-scientific ridiculousness. This time, they\u2019ve published a paper indicating that a simple facial recognition algorithm can tell a person\u2019s political affiliation.<\/p>\n

First things first: The paper is called \u201cFacial recognition technology can expose political orientation from naturalistic facial images.\u201d You can read it here<\/a>. Here\u2019s a bit from the abstract:<\/p>\n

\n

Ubiquitous facial recognition technology can expose individuals\u2019 political orientation, as faces of liberals and conservatives consistently differ.<\/p>\n<\/blockquote>\n

Second things second: These are demonstrably false<\/em> statements. Before we even entertain this paper, I want to make it completely clear that there\u2019s absolutely no merit to Kosinski and his team\u2019s ideas here. Facial recognition technology cannot expose individuals\u2019 political orientation.<\/p>\n

[Related: The Stanford gaydar AI is hogwash<\/a>]<\/em><\/p>\n

For the sake of brevity I\u2019ll sum up my objection in a simple statement: I once knew someone who was a liberal and then they became a conservative.<\/p>\n

While that\u2019s not exactly mind blowing, the point is that political orientation is a fluid concept. No two people tend to \u201corient\u201d toward a specific political ideology the same way.<\/p>\n

Also, some people don\u2019t give a shit about politics, others have no clue what they\u2019re actually supporting, and still others believe they agree with one party but, in their ignorance, don\u2019t realize they actually support the ideals of a different one.<\/p>\n

Furthermore: since we know the human face doesn\u2019t have the ability to reconfigure itself like the creature from \u201cThe Thing,\u201d we know that we don\u2019t suddenly get liberal face<\/i> if one of us decides to stop supporting Donald Trump and start supporting Joe Biden. <\/span><\/p>\n

This means the researchers are claiming that liberals and conservatives express, carry, or hold themselves differently. Or they\u2019re saying you\u2019re born a liberal or conservative and there\u2019s nothing you can do about it. Both statements are almost too stupid to consider.<\/span><\/p>\n

The study claims that demographics (white people are more likely to be conservative) and labels (given by humans) were determining factors in how the AI segregates people.<\/span><\/p>\n

In other words, the team starts with the same undeniably false premise as many comedians: that there\u2019s <\/span>only two kinds of people in the world<\/i>. <\/span><\/p>\n

According to the Stanford team, the AI can determine political affiliation with greater than 70% accuracy, which is better than chance or human prediction (both being about 55% accurate). <\/span><\/p>\n

Here\u2019s a an analogy for how you should interpret the Stanford team\u2019s claims of accuracy: I can predict with 100% accuracy how many lemons in a lemon tree are aliens from another planet. <\/span><\/p>\n

Because I\u2019m the only person who can see the aliens in the lemons, I\u2019m what you call a \u201cdatabase.\u201d If you wanted to train an AI to see the aliens in the lemons, you\u2019d need to give your AI access to me.<\/span><\/p>\n

I could stand there, next to your AI, and point at all the lemons that have aliens in them. The AI would take notes, beep out the AI equivalent of \u201cmm hmm, mm hmm\u201d and start figuring out what it is about the lemons I\u2019m pointing at that makes me think there\u2019s aliens in them. <\/span><\/p>\n

Eventually the AI would look at a new lemon tree and try to guess which lemons <\/span>I <\/i>would <\/span>think<\/i> have lemons in them. If it were 70% accurate at guessing which lemons <\/span>I<\/i> <\/span>think<\/i> have lemons in them, it would still be 0% accurate at <\/span>determining<\/i> which lemons have aliens in them. Because <\/span>lemons don\u2019t have aliens in them. <\/i><\/p>\n

That, readers, is what the Stanford team has done here and with its silly gaydar. They\u2019ve taught an AI to make inferences that don\u2019t exist because (this is the important part): there\u2019s no definable scientifically-measurable attribute for political party. Or queerness. One cannot <\/span>measure<\/i> liberalness or conservativeness because, like gayness, there is no definable threshold. <\/span><\/p>\n

Let\u2019s do gayness first so you can appreciate how stupid it is to say that a person\u2019s facial makeup or expression can determine such intimate details about a person\u2019s core being.<\/span><\/p>\n

    \n
  1. If you\u2019ve never had sex with a member of the same sex are you gay? There are \u201cstraight\u201d people who\u2019ve never had sex.<\/span><\/li>\n
  2. If you\u2019re not romantically attracted to members of the same sex are you gay? There are \u201cstraight\u201d people who\u2019ve never been romantically attracted to members of the opposite sex.<\/span><\/li>\n
  3. If you used to be gay but stopped, are you straight or gay?<\/span><\/li>\n
  4. If you used to be straight but stopped, are you straight or gay?<\/span><\/li>\n
  5. Who is the governing body that determines if you\u2019re straight or gay?<\/span><\/li>\n
  6. If you have romantic relations and sex with members of the same sex but you tell people you\u2019re straight are you gay or straight?<\/span><\/li>\n
  7. Do bisexuals, asexuals, pansexuals, demisexuals, gay-for-pay, straight-for-a-date, or just generally confused people exist? Who tells them whether they\u2019re gay or straight? <\/span><\/li>\n<\/ol>\n

    As you can see, queerness isn\u2019t a rational commodity like \u201cenergy\u201d or \u201cnumber of apples on that table over there.\u201d <\/span><\/p>\n

    The Stanford team used \u201cground truth\u201d as a measure of gayness by comparing pictures of people who said \u201cI\u2019m gay\u201d to pictures of people who said \u201cI\u2019m straight\u201d and then fiddled with the AI\u2018s parameters (like tuning in an old radio signal) until they got the highest possible accuracy. <\/span><\/p>\n

    Think of it like this: I show you sheet of portraits and say \u201cpoint to the ones that like World of Warcraft.\u201d When you\u2019re done, if you didn\u2019t guess better than pure chance or the human sitting next to you I say \u201cnope, try again.\u201d <\/span><\/p>\n

    This goes on for thousands and thousands of tries until one day I exclaim \u201ceureka!\u201d when you manage to finally get it right. <\/span><\/p>\n

    You have not learned how to tell World of Warcraft players from their portraits, you\u2019ve merely learned to get that sheet right. When the next sheet comes along, you\u2019ve got a literal 50\/50 chance of guessing correctly whether a person in any given portrait is a WoW player or not.<\/span><\/p>\n

    The Stanford team can\u2019t define queerness or political orientation like cat-ness. You can say that\u2019s a cat<\/em> and that\u2019s a dog<\/em> because we can objectively define the nature of exactly<\/em> what a cat is. The only way you can determine whether someone is gay, straight, liberal, or conservative is to ask them. Otherwise you\u2019re merely observing how they <\/span>look<\/i> and <\/span>act<\/i> and deciding whether <\/span>you<\/i> believe they are liberal or queer or whatnot. <\/span><\/p>\n

    The Stanford team is asking an AI to do something <\/span>no human<\/i> can do \u2013 namely, predict someone\u2019s political affiliation or sexual orientation based on the way they look. <\/span><\/p>\n

    The bottom line here is that these silly little systems use basic algorithms and neural network technology from half-a-decade ago. They\u2019re not smart, they\u2019re just perverting the literal same technology used to determine if something\u2019s a hotdog or not<\/a>.<\/span><\/p>\n

    There is no positive use-case for this.<\/span><\/p>\n

    Worse, the authors seem to be drinking their own Kool Aid. They admit their work is dangerous, but they don\u2019t seem to understand why. Per this Tech Crunch article<\/a>, Kosinski (referring to the gaydar study) says:<\/p>\n

    \n

    We were really disturbed by these results and spent much time considering whether they should be made public at all. We did not want to enable the very risks that we are warning against. The ability to control when and to whom to reveal one\u2019s sexual orientation is crucial not only for one\u2019s well-being, but also for one\u2019s safety.<\/p>\n

    We felt that there is an urgent need to make policymakers and LGBTQ communities aware of the risks that they are facing. We did not create a privacy-invading tool, but rather showed that basic and widely used methods pose serious privacy threats.<\/p>\n<\/blockquote>\n

    No, the results aren\u2019t scary because they can out queers. They\u2019re dangerous because they could be misused by people who believe <\/em>they can. Predictive policing<\/a> isn\u2019t dangerous because it works, it\u2019s dangerous because it doesn\u2019t work: it simply excuses historical policing patterns. And this latest piece of silly AI development from the Stanford team isn\u2019t dangerous because it can determine your political affiliation. It\u2019s dangerous because people might believe it can, and there\u2019s no good use for a system designed to breach someone\u2019s core ideological privacy, whether it works or not.<\/p>\n

    Published January 14, 2021 \u2014 20:41 UTC <\/p>\n

    Source<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"

    Stanford researcher Michael Kosinski, the PhD behind the infamous \u201cGaydar\u201d AI, is back with another phrenology-adjacent (his team swears it\u2019s not phrenology) bit of pseudo-scientific ridiculousness. This time, they\u2019ve published a paper…<\/p>\n","protected":false},"author":1,"featured_media":2313,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[1],"tags":[],"_links":{"self":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/posts\/2312"}],"collection":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=2312"}],"version-history":[{"count":0,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/posts\/2312\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=\/wp\/v2\/media\/2313"}],"wp:attachment":[{"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=2312"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=2312"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.londonchiropracter.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=2312"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}