Leaf&Core

Google Appoints Anti-Immigrant Transphobe to AI Ethics Board

Danielle from Leaf and Core
Reading Time: 7 minutes.
Google logo in front of light purple background.

Photo: Nick Statt / The Verge

There are things that just don’t belong together. One clear example is that of bigots and ethics committees. They mix like oil and water. Hate doesn’t have a place in ethics. Google disagrees.

Google created an external ethics board to monitor their use of AI. This should help Google avoid bias in their machine learning algorithms. Though these algorithms already seem to show some bias. They’re “hidden parameters,” biases of humans reflected in AI, even though they weren’t part of the core data set.

For example, AI can have biases towards any minority group. Facial recognition software is famously less likely to identify people of color or women. AI used to drive autonomous vehicles is more likely to hit a black person than a white person, because it’s less likely to identify black people as people. No, really. The software could also be transphobic, labeling a trans woman as a man or a trans man as a woman, or identifying a non-binary person as a binary gender.

These are exactly the things an ethics advisory board should put a stop to. However, Google brought on a potentially problematic person, the president of the far-right, anti-immigrant, and anti-LGBTQ Heritage Foundation, Kay Cole James. She has spoken critically of LGBTQ rights, especially trans rights, and would not hold Google to a high standard of integrity. In fact, she may permit or boost Google’s existing biases against LGBTQ people.

What is Google’s AI Ethics Board?

Last year, Google revealed their AI principles. These were their goals for AI, as well as their guidelines for keeping AI ethical. They were:

  1. Be socially beneficial
  2. Avoid creating or reinforcing unfair bias
  3. Be built and tested for safety
  4. Be accountable to people
  5. Incorporate privacy design principles
  6. Uphold high standards of scientific excellence
  7. Be made available for uses that accord with these principles.

AI can easily violate these principles. It could be used for war machines, to incorrectly identify a person in a way that could put them in danger, reveal their private information by identifying individuals through their travels or habits, reinforce biases and obscure methods, or be sold to the highest bidder.

Google has already violated a number of these principles. That’s why they introduced an external AI ethics board, to help them stay in line, or, at least, appear as though they’re attempting to stay in line. Google sourced their committee from experts in machine learning, math, natural language processing, music, industrial engineering, psychology, philosophy and ethics, a diplomat, and someone who’s president of a hate group calling itself a think tank.

One of those things is not like the others.

Who is Kay Cole James?

 

Kay Cole James has spoken out against transgender people, misgendering them and pushing against equal rights, despite what all health professionals agree on: transgender people are valid.

Kay Cole James is currently the president of the Heritage Foundation, a conservative think tank most well known for their work against LGBTQ and women’s rights. She worked under the administrations of both Bush presidents. If you ignore her beliefs, she’s had an incredibly respectable career. However, when we’re talking about ethics, we can’t ignore her beliefs.

James called Mexican and Central American immigrants and asylum seekers “dangerous criminals, drug smugglers and sex traffickers.” She called trans women men and made the claim that including transgender women in the category of women can “erase efforts to empower women economically, socially, and politically,” though she made no efforts to back her claims up or support them with factual, statistical, or even anecdotal evidence. That is the very nature of an act of bigotry, blind fury and hate towards a group for no reason. She spoke out in favor of businesses having the right to discriminate, just as they did in the 1950s, except now against LGBTQ people and unwed mothers, instead of people of color. She has even defended coal. Coal! One of the least efficient and most harmful methods of generating power.

She’s not only a bigot, she’s a dangerous one. She’s pushing for policies that will hurt women, transgender people, all LGBTQ people, and, literally everyone on the planet. What place could she possibly have on a board dedicated to ethics? Since when has endangering every life on the planet been an ethical decision?

What is Her Position on Google’s New AI Ethics Board?

Google’s own employees protesting the company. Photo: James Martin/cnet

 

Why on earth would Google consider this hateful women for a position on an ethics board? To consider that, we have to look into the fallacy of “both sides” or “enlightened centrism” thinking, as well as the reaction of conservative groups to other decisions to limit hate speech and harassment.

The Fallacy of Representing Both Sides as “Fair”

It’s reasonable to say you should hear both sides of an argument. However, it’s a fallacy to believe that both sides of the argument come from level ground. That is, it’s foolish to think they are based on the same facts, empathy, or morality.

For example, let’s say you witness two students arguing on a playground. Remaining impartial, you decide to hear both sides of the argument. Good so far. The first person holds out a piece of paper. On it is the letter ‘A.’ They say, “The letter ‘A’ is written on this piece of paper.” An English teacher, literature teacher, principal, and 10 students sit around and agree. The paper can be seen by all, you can see the letter ‘A’ on it. The other person picks up a rock and throws it at the first person’s head, exclaiming, “How could you trust someone with blood coming down from their forehead?”

Are both sides of this argument rational, logical, based in science, observation, or expert opinion? Do you step back and nod your head, stating that they have equivalent but opposing views? No. In fact, one is both an ad hominem and a physical attack.

That’s a Silly Example

You’re right, it is! But it illustrates the fallacy of going in, exclaiming that both sides of an argument are equivalent just because there are two sides of an argument. We have mountains of proof and over 98% of scientists agreeing, climate change is real, the earth is getting hotter, and it is humanity that has caused it through the rampant burning of fossil fuels. The other side of this argument is not based in science, observation, factual data, and often comes from people with financial ties to the fossil fuel industry.

They are not equivalent beliefs.

The same can be said of LGBTQ people. As James has made transgender people a primary target of her anti-equality campaign, let’s consider gender theory. People who are transgender all report the same feelings. Their doctors recognize their gender as valid. Psychologists and physicians agree, transgender people are the gender they claim to be. Even MRI scans validate these positions. However, with no evidence or facts, the right claims that these people, contrary to the beliefs of medical professionals, are wrong, that they do not know their genders, that they are not valid, and that they are actually dangerous. Again, none of this is rooted in any evidence.

The arguments are far from equal. One is based in facts, studies, and medical professionals. The other is baseless hate.

Okay, Those Are Better Examples. Where Does Google Come In?

From the Google employee walkout in 2018. Photo: Stephen Lam/Reuters

 

This is why Google brought Kay Cole James on their ethics board. She’s a “conservative voice,” that is, someone who brings a conservative viewpoint.

Now, I’m not saying all conservative voices are hateful. Not all conservatives are anti-immigrant or anti-LGBTQ. These are the conservative voices Google and others should elevate, those who believe local taxes are more vital than federal, or those who argue in favor of defense spending. Though overzealous now, defense is important for any nation. When AI is concerned, we do have to think about the ways it could be used to defend the country, humanity, or our values. A person can be a conservative without being a hateful bigot.

But bringing on a conservative voice does not mean you must throw in with avowed homophobes, transphobes, or xenophobes. One does not have to align with evil to affect the greater good.

Warding Off Republicans and the “Alt-Right”

Mark Zuckerberg speaking to congress. Some Republicans wanted to know if Facebook held bias. Other tech leaders were similarly forced to speak. Photo: JIM WATSON/AFP/Getty Images

 

There’s another reason Google brought her on. Google, Twitter, and other companies have suffered abuse at the hands of conservative politicians. Because they’ve filtered out hate speech and harassment, which overwhelmingly is more prevalent among conservatives, Republicans have accused Twitter, Facebook, and Google of bias. This, again, falls back to the idea that not all positions are equal. A person trying to live their life and talk about their hobbies or job on Twitter has a greater right to do so than a person who wants to harass, assault, and attack people because they’re different.

Rights extend as far as their influence. If a person’s actions cause harm, they’re no longer protected rights. Nor are they valid positions.

Google likely gave James her role in an effort to quell these complaints before they come in. However, once again, they could have chosen a less hateful conservative voice.

Can a Bigot Contribute to an Ethics Board?

Simply put? No, they can’t. Take arguments for “free speech.” We see this on college campuses and from Republican candidates frequently. They blame liberals for attacking free speech. But we all love free speech, don’t we? How could liberals do this? Free speech and the free press is an ethical idea, the basis of a democracy!

What they don’t mention is that the conservatives coming to campus were neo-Nazis, members of other hate groups, or singled out individual students for harassment. They don’t mention the conservatives harassing women by shouting in their faces as they walk into a health clinic, shooting wildly into a crowd, chasing down people of color, or running over counter-protestors in cars. Their “free speech” defense is really in favor of hate speech, which silences minorities, silences victims, and incites violence. Suddenly their seemingly innocent “free speech” platform is not innocent at all. It’s not ethical, and advocates for hate speech that incites harm.

An example of a fringe and hateful conservative position. Hate speech on a Facebook post. Facebook allowed it to keep from upsetting conservatives, fueling anti-Muslim violence.

 

Can a conservative be ethical? Yes. Can a bigot be ethical? By their very nature, no. Because of that, people like Kay Cole James have no place on an ethics committee.

What’s the Harm?

This risk assessment was wrong, and damaged lives as a result. AI caused this harm because the people who made it didn’t realize it was racially biased.

Let’s say a piece of software is more likely to classify a transgender woman as a male, and Google is working with a contractor to use it in airport screenings. That same software could lead to the woman being pulled out of line, mistreated, or harassed. In some countries, she could be arrested, beaten, or even killed. Even in the United States, it could put her life at risk.

Most people would not approve of this use of AI. However, someone who hates transgender people would see no problem.

Consider software that is more likely to misidentify people of color, as facial recognition often does. Installing it in a place that automatically would alert immigration would be problematic. It would often flag innocent people for harassment from ICE. Furthermore, depending on where it’s installed, it could lead to harm or death. If such facial recognition is installed at hospitals, people of color and undocumented immigrants would be less likely to seek help.

Someone who doesn’t care about undocumented immigrants may not see a problem with this.

So what’s the harm? If she’s a deciding vote on any matter, it will skew towards harming LGBTQ people and undocumented immigrants. Just one person can skew that bias. That one vote could bring harm to someone.

Is it worth it just to keep Republicans off your back? I don’t think so. Trading someone’s safety for convenience is never the right thing to do.

Does Google Care?

Living Hope’s hateful anti-LGBTQ message. The app was allowed on Google Play.

 

It’s hard to say. This isn’t the first time even this month that Google has come under fire. They also refused to remove an anti-LGBTQ conversion therapy app from the Google Play store. However, Google finally removed the app. It wasn’t because of the fact that conversion therapy increases self harm and suicide rates among those who go through it, but because the Human Rights Campaign gave Google a very public slap on the wrist that will sting them for an entire year.

It seems Google doesn’t care about doing harm to LGBTQ people. Their actions regarding the Google Play store and now the appointment of someone who very publicly practices hate speech about transgender people proves that. The question is whether or not Google is passively uncaring, or actively targeting LGBTQ people. With the decision to allow a hateful app to remain on their store, Google seemed to play a passive role in spreading hate. Now, they’ve made an active decision to elevate it.

P.S.

Today is Transgender Day of Visibility. Just as Google defended Absher, an app that allows men to control women in Saudi Arabia, during the week leading up to International Women’s Day, their timing here could not be worse.

Sources:
Exit mobile version