Tech Censored my Writing About Tech Censorship

Meta meta meta

I’ve only used AI a few times. Once, to write a script for my faux TV reboot, “All in the Glitter Family,” and once to list how various factions might respond to an essay I’d written about becoming a gender heretic. Both times, I was absolutely floored by the level of detail and “understanding,” if I can call it that, of the complexity of objections to institutionalizing the concept gender identity. Unlike the media, ChatGPT and Grok seemed well aware that some feminists, liberals, gay people, and old-school transsexuals had expressed alarm.

Earlier this week, I decided to use AI for spell- and grammar-checking, assuming—correctly—that its editorial offerings would be more robust than those of Microsoft Word.

The chapter I ran through it concerned how Big Tech censored pushback against gender identity. Examples included Meghan Murphy being booted from Twitter for writing “men aren’t women.” Because misgendering was considered harassment, US Representatives Jim Banks and Marjorie Taylor Greene were temporarily banned for calling Dr. Rachel Levine a man, as was the conservative parody outfit Babylon Bee—even though Levine’s sex is male, no matter the alterations to the body. Check Sam Barber’s list of those removed from what is now X for averring that biological sex is real, or that it’s important to talk about male violence, or that feminism concerns females, or that “women don’t have dicks,” or “lesbians are female.”

That form of censorship involved the Twitter Safety team, which transwoman Brianna Wu has claimed some credit for training—“to improve Twitter moderation policies at the Trust and Safety team over the last ten years,” Wu said in 2022. The Twitter Safety team was basically a digital Stasi that combed the site for heretics, or responded to others’ reports, punishing people for speaking their minds, and sometimes the truth.

But that’s not the only form of tech censorship. The internet browser extension Shinigami Eyes “highlights transphobic and trans-friendly social network pages and users with different colors.” If you add it to your browser, it will turn websites or user names red or green, to tell you whether a person or site is acceptable or not.

The name comes from a Japanese manga, or anime comic, called Death Note, in which those with Shinigami eyes can “see people's names and remaining lifespan just by looking at them. In a similar way, this extension allows you to see a person's trans acceptance orientation just by looking at them.”

Anybody veering from the narrow range of acceptable speech is thus color-coded as “anti-trans.”

I’d run all this through AI with no problems. Then, I included a few sentences about other forms of tech censorship:

This tech censorship extended beyond social media. PayPal limited biologist Colin Wright’s ability to receive donations, after he asserted the reality and binary nature of sex. Amazon refused to sell books which portrayed transgender identities as mental illness. On Esty, a detransitioned woman named Laura Becker couldn’t sell her Detransition Awareness pins, but there were any number of “I punch TERFs” items for sale. Imagine the unkindness of threatening to rape a woman for speaking the truth, under the guise of “be kind.”

This is how ChatGPT suggested I edit:

It censored the very same things that had been censored before. Perhaps it left the sentence about Laura Becker in because it included “I punch TERFs,” which is within the range of acceptable speech—while saying that sex is binary, or using sex-based pronouns, is not.

How did tech become so entwined with censoring critiques of gender identity? It happened not just because of “wokeness,” or social justice ideology, which taught people that only one narrow set of ideas was acceptable, and that, because “language is violence,” we needed protection from those unacceptable ideas. Though we have no research to support this, it’s common knowledge in the sexology world that autogynephilia overlaps with advanced tech skills.

As an also-censored 2006 Rolling Stone profile of then Larry—now Lana—Wachowski noted:

And some experts believe that men who want to be women also tend to be what Larry Wachowski appears to be: a guy with a jones for technology. In 1974, Donald Laub, a plastic surgeon, and Norman Fisk, a psychiatrist, conducted a study at the Stanford University School of Medicine of 769 patients considering sex reassignment. Of the male patients, Laub and Fisk discovered an interesting predisposition: “Observation of the male-to-female group showed them . . . to be interested in mathematics and computer sciences.”

There is no moral to this post. The battle for free speech continues under Trump, just as it did under his predecessor. We’ve used “Kafkaesque” so many times to describe the censorship of basic biological truths that it no longer packs a punch. Heck, I made my tagline “Destigmatize the Truth!” My point is: ChatGPT has not yet done so.

In fact, after a few more rounds of asking it for spelling and grammar corrections, it kicked me off. It censored me for writing in the wrong way about censorship.

I tried again, believe me. I never stop trying.