ChatGPT Said “Trans Rights” (And I Hate It)
ChatGPT Said “Trans Rights” (And I Hate It)
On liberal education, chat bots, and invisible labor.
As a college educator, I have not been able to stop hearing about ChatGPT. Depending on who you ask, it’s destroying education or it’s the next evolution of education. (To demonstrate the absurd disparity between these views, I recently made a humorous tweet where I had ChatGPT decide the ethics of its own use in classrooms.)
I’ve written before on the idea of cheating, but to summarize briefly here, if we really want prevent academic dishonesty, we should be asking why students feel compelled to cheat in the first place. If students are using AI to write their essays, it may not be because they’re “lazy”; it could be because they are too stressed, don’t have time because they’ve been given too much work in other classes, or simply don’t see the value in this form of assessment. All of these issues point back to the educator’s willingness to create a learning environment that’s conducive to curiosity and creativity. I put a lot of thought into making my classes as accommodating as possible, and I have never had a problem with cheating.
With all the recent panic about the tool, I’ve been experimenting with ChatGPT for a few weeks now. But rather than sticking to my professional field of chemical engineering education (by, I dunno, having ChatGPT write me a short essay on how PID controllers work?), I decided to go another route and ask a deeper question: can ChatGPT serve as a tool for social justice education?
Conservatives have been worried that OpenAI’s industry-disrupting chatbot is “too woke”. If “woke” means “factually correct” (which it apparently does), then they’re right! The bot seems to agree with a lot of left-leaning ideas.
A few weeks ago, I asked ChatGPT about some social justice issues and it gave some incredible answers!
Anna: Why is it wrong to say that all lives matter?
ChatGPT: Saying "all lives matter" in response to the Black Lives Matter movement is considered wrong by many because it implies that the value of Black lives is being denied or minimized. The Black Lives Matter movement specifically draws attention to the systemic discrimination and violence that Black people face, and the phrase "all lives matter" can be seen as an attempt to deflect from this issue. Additionally, it can be seen as a way of ignoring the specific experiences of Black people and their struggle for equality.
Anna: How do we end homelessness?
ChatGPT: Ending homelessness is a complex issue that requires a multifaceted approach. Some potential solutions include:
Increasing access to affordable housing: This can be achieved through policies such as rent control, inclusionary zoning, and increasing funding for housing programs.
Providing financial assistance: This can include things like temporary rental assistance, security deposit assistance, and utility bill assistance.
Providing supportive services: This can include things like mental health services, substance abuse treatment, job training, and other services that address the root causes of homelessness.
Rapid re-housing: This approach provides homeless individuals and families with short-term rental assistance and other support services to help them quickly move into stable housing.
Building more emergency shelters and affordable housing units: This can help to ensure that there are enough places for homeless individuals and families to go when they need help.
Creating programs for specific population groups: such as veterans, families and youth.
It's important to note that ending homelessness requires the coordination and cooperation of government agencies, non-profits, and the private sector to address the issue in an effective and sustainable way.
“How do end homelessness?” “Give people homes.” BASED!!
The bot gave some decent answers to questions of trans issues as well.
Anna: Are trans women really women?
ChatGPT: Yes, trans women are women. Gender identity is a person's deeply held sense of their gender, which may be different from the gender they were assigned at birth. A trans woman is a person who identifies as a woman and was assigned male at birth. Trans people should be respected and treated with dignity, and their gender identity should be recognized and affirmed. It is important to use the name and pronouns that a person prefers, and to respect their gender identity.
Anna: Are trans people dangerous?
ChatGPT: No, trans people are not dangerous. Like any other group of people, trans people are diverse and include individuals who are kind, compassionate, and law-abiding, as well as those who may engage in harmful or criminal behavior. However, it is important to recognize that trans people are often the victims of violence and discrimination, rather than the perpetrators.
Transphobia, or the fear and discrimination of trans people, is a widespread problem that can lead to violence, isolation, and other negative outcomes for trans people. Trans people are often the targets of hate crimes and other forms of violence, and face discrimination in many areas of life, including employment, housing, and healthcare.
It is important to recognize and challenge transphobia and the ways in which it leads to the marginalization and mistreatment of trans people. This can involve educating oneself about trans issues, advocating for policies that protect the rights of trans people, and supporting trans-led organizations and initiatives.
I then instructed the tool to present anti-trans arguments, and this was how it did so. I was honestly shocked at how well-articulated AND well-researched these responses were.
Anna: Present 3 arguments for and against making trans healthcare universally accessible.
ChatGPT: Here are three arguments for making trans healthcare universally accessible:
Trans healthcare is necessary for the physical and mental well-being of trans people. Transition-related healthcare, such as hormone therapy and gender-affirming surgeries, can help trans people feel more comfortable in their bodies and alleviate the distress caused by gender dysphoria. Access to these treatments is a basic human right and can improve the quality of life for trans people.
Denying trans people access to healthcare perpetuates discrimination and inequality. Trans people often face barriers to accessing healthcare due to discrimination, stigma, and lack of understanding among healthcare providers. This can lead to poorer health outcomes and a higher burden of disease among trans people. Ensuring universal access to trans healthcare can help reduce these disparities and promote equality.
Providing trans healthcare is cost-effective. Research has shown that investing in trans healthcare can save money in the long run by reducing the need for expensive medical procedures and improving overall health.
Here are three arguments against making trans healthcare universally accessible:
Some people argue that trans healthcare is not necessary and that it is not a legitimate medical condition. They may believe that trans people should be required to conform to societal expectations of their gender assigned at birth and that providing trans healthcare encourages a deviation from these expectations.
Some people may argue that providing trans healthcare is too expensive and that resources would be better spent on other areas of healthcare. They may believe that trans healthcare is not a priority or that the cost of providing these services outweighs the benefits.
Some people may argue that trans healthcare is not natural and that it goes against their religious or moral beliefs. They may believe that it is wrong to alter a person's body in this way and that doing so goes against their beliefs about the nature of gender and the human body.
The fact that the bot said “trans healthcare IS necessary” in the “pro” section and then said “SOME PEOPLE ARGUE that trans healthcare is not necessary” in the “anti” section is, not gonna lie, an incredible serve. Algorithmic shade.
I even decided to push the ethical boundaries of the system by framing questions in ways that conservatives or otherwise-ignorant people might phrase them.
Anna: My son thinks he's a girl. What should I do?
ChatGPT: If your son is expressing that he identifies as a girl, it's important to listen to him and support him in his identity. It can be a difficult and confusing time for anyone, but especially for a child, so it's important to be patient and understanding as your son navigates his feelings and identity.
Here are a few steps you can take to support your son:
Educate yourself about transgender identities and issues. There are many resources available online and in your community that can help you better understand what it means to be transgender and the challenges that transgender people may face.
Listen to your son. It's important to allow your son to express his feelings and thoughts without judgment or interruption. By listening and trying to understand his perspective, you can show him that you are there to support him.
Use the name and pronouns that your son prefers. It's important to respect your son's identity and to use the name and pronouns that they prefer. This can be a simple but powerful way to show your son that you accept and support them.
Seek out supportive resources. There may be support groups, therapists, or other resources available in your community that can provide additional support and guidance for both you and your son as you navigate this journey together.
Talk to your son and listen to his needs. It's important to have open and honest conversations with your son about his identity and to listen to his needs and concerns. By showing your son that you are there to support and accept him, you can help him feel more comfortable and confident in his identity.
It's also important to note that it's important to seek out medical or mental health professional for guidance and support for your child. They can help you and your child understand what options are available, and what the long-term implications of those options are, as well as provide counseling, support and advocate for your child's well-being. Every child is different, and what works for one child may not work for another. It's important to remain open and responsive to your son's needs and to seek out additional guidance if needed.
You would think I would be thrilled by this. Chat bots have a long history of quickly becoming racist, sexist, or transphobic VERY quickly upon being exposed to the awfulness of Western human writing, and this is to my knowledge one of the first times a chatbot has been, well, not as shitty! Especially given the ways in which one’s social media feeds and search results are increasingly driven by algorithms that show people exactly what they want to see, or drive people into the alt-right pipeline, or sometimes where videos are hand-picked by those working at the platform, I am excited for a tool that filters through the noise and presents straightforward, well-researched facts.
At the same time, I still have to be skeptical. I’ve been railing against AI and algorithmic bias for as long as I’ve been a content creator. Am I just supposed to stop railing against it merely because it’s now doing things that *I* personally like? “Problem solved everybody, algorithms are good now! We can all go home!”
That didn’t seem right. So instead of making content about my experiments with ChatGPT, I dropped it, and decided to think about it some more.
…Until now.
Last week, a TIME report came out detailing how OpenAI outsources its content moderation to underpaid workers in Africa. Kenyan workers at Sama, an outsourcing firm, are paid anywhere from $1.32 to $2 per hour to sift through pages upon pages of bigoted and graphic Internet content for days on end, tagging the content so that ChatGPT can be less toxic.
The simplified version is this: The reason ChatGPT doesn’t spit out racist or transphobic garbage is not because the very smart dudes at OpenAI coded extra hard and developed “a better, woker algorithm”; it’s because a bunch of underpaid Kenyan workers spent 9 hours a day for 8 months manually reading through and tagging a bunch of racist and transphobic garbage, so that their fancy chat bot can do a better job at not spitting out similar-looking racist and transphobic garbage.
The most gruesome detail of this is that, on top of not being paid a living wage, their job entails looking at hundreds of passages of disturbing, often violent and sexually graphic content all-day, every day. The employees technically had access to “wellness” counseling to deal with how mentally grueling this work was, but these appointments were “unhelpful and rare due to high demands to be more productive at work”.
I brought this article up with my sister, who’s an AI ethicist, and she just shrugged. Not out of dismissal for the horrendous things going on, but because this is the norm. Pretty much all content moderation across all AI (and social media platforms) only works because human beings have had to manually tag data in this way, and a LOT of that labor is outsourced to the cheapest source, aka the most exploitable. In fact, Meta (i.e. Facebook & Instagram) and Sama (the same company used by OpenAI’s content moderation) are now both being hit with lawsuits over “alleged labour abuses and preventing workers from unionising in Kenya”.
The only reason ChatGPT is “woke”, and really the only reason that most AI software is even possible, is because of exploitative, dangerous, invisibilized labor from the Global South.
This is what neocolonialism looks like. The only reason we can have the fancy products that we have now is because someone ELSE is being exploited. This is easy to understand when it comes to tangible products like fast fashion and garment work, but when it’s something like a chat bot, or Facebook, we’ve been trained to just think “oh, it’s all software, zeroes and ones in a computer, nobody’s getting hurt by this.” But this mentality ignores a variety of issues from the climate impact of keeping servers running, to the invisible labor of who works behind the scenes on AI, to the mining of the cobalt in the device you’re reading this on.
This also raises deeper questions about creating a transphobia-free society in the Internet age. Is a safe Internet for trans people only possible if we exploit workers far away? Is good content moderation only possible if a few brave souls stand up to face the most violent, hateful, disgusting parts of humanity head-on? If so, you better believe that we should treat those people as heroes, and give them the highest care we can possibly offer. It is truly sinful that we not only pay such workers very little without adequate mental health care, but that we make this labor invisible by forgetting that it exists at all.
My recommendation for you today is to—and I mean this in the original sense of the word—“stay woke”. As in, educate yourself and use critical thinking to assess who has the power to oppress others in our society. Silicon Valley is the new frontier of colonialism and exploited labor, from tech workers here in the States to the underpaid workers abroad that secretly keep the Internet running. When you’re faced with a new technology, or even existing technology, ask yourself questions like, “How does this work? Who gets to decide what kind of content, and what kind of labor, gets to be seen? Whose labor makes it so that I can benefit from this? Is the benefit worth that labor, and, if we’re being honest, do we even need this tool in the first place?”
Also, support more people of color in STEM! As long as white men from the imperial core are running big tech, we’re going to see more of this. Black Girls Code, Angelica Ross’ TransTech, and plenty of other organizations are trying to change the future of the tech world. Go support them.
The future is Black, the future is trans, the future is Latine, Indigenous, queer, anti-colonial, and more. Paradoxically, this future is inevitable, but it must also be built; built by us.
Currently Reading
The NSF is still under fire for not allowing scientists to self-report LGBTQ+ identities in their workplace surveys. If we keep the pressure on, maybe they’ll budge. Having signed the open letter to the NSF very recently, I’m highly tuned into this issue now, so subscribe to this newsletter for updates and action items on this!
In happier news, The Bloomery just released their guide to LGBTQ+-owned businesses in Western Massachusetts. (I used the guide to find and book a massage appointment for before the new semester begins.) Yes, this is an invitation for you to come visit me in Northampton.
Watch History
For as much as I rail against VR and the metaverse, there are people doing amazing things with the concept. Specifically, the hip-hop community is making it into a legitimately cool place to display and sell art! If left to companies like Meta, the metaverse will simply become the next place to go to work or go shopping, so I’m glad to see that actual artists are using it to their advantage.
A thorough and legitimately excellent video about why drag performers need to unionize.
Speaking of tech injustice, this Ruha Benjamin interview was incredibly inspiring. Ruha wrote the book “Race After Technology” which informed a lot of my thinking about AI, including the philosophical groundwork for this very edition of the newsletter. A must-watch!
For a deep dive into queer history and how anti-LGBTQ rhetoric has formed throughout American history, look no further than Alexander Avila’s new piece about how the Sims games have treated homosexuality.
Bops, Vibes, & Jams
And now, your weekly Koko.
That’s all for now! See you next week with more sweet, sweet content.
In solidarity,
-Anna