The Unbearable White Maleness of AI

We have entered the era of the cute “AI” stunt, and its implications are more immediately disconcerting than the looming specter of a robot apocalypse (and certainly more amusing). The gag goes something like this: A journalist, tasked with covering “artificial intelligence,” asks a computer program to do something for them, such as write a paragraph or two, draw a silly picture, make restaurant recommendations, or answer a question about itself. Thus edified, the writer makes pronouncements about the future of AI and possibly the future of humanity.

The success of the stunt hinges on not looking too closely at whether “artificial intelligence” actually exists (it doesn’t, and won’t for a long time, if ever, but that semantic ship has sailed). It’s a lot more fun to pretend like you’ve “handed [your] autonomy” over to a computer for a day than it is to acknowledge that going wherever ChatGPT sends you for breakfast is no more interesting or futuristic than Googling “good pancakes nearby” and heading wherever the “I’m Feeling Lucky” button lands you.

Men—especially white dudes, bless their hearts—seem particularly susceptible to the conceit of so-called “AI,” even ones who otherwise seem pretty smart (and Elon Musk). Nothing gets to the heart of this better than comedian Avery Edison’s very good joke: “I wrote ‘I am alive’ on a piece of paper, and placed it into a photocopier. What I saw next has shocking implications.” 

This is where we’re at: A bunch of guys hollering “It’s aliiiiiiiiiiive!” whenever they send ChatGPT-generated emails because some other guys programmed a powerful autocorrect software to issue the results of mathematical calculations in the first person. Whether they’re hollering out of terror or excitement, both reactions assume the same posture, defaulting to the presumption that “AI” must either be conquered or capitulated to. It’s why we have AI’s most prominent detractors calling for a pause in development to avert the apocalypse, and AI’s thirstiest hypesters arguing that it’ll be the end of the world if we do stop now.

Win and live or lose and die. Control, never cooperate. It’s no accident that the word “robot” is derived from a Slavic term for “forced labor,” or that our digital assistants Siri and Alexa have feminine names and voices, or that some facial recognition software literally cannot identify people of color. Ensuring that technological advancements serve the most privileged at the expense of everyone else is a core principle of empire, and thus of digital colonialism, and what some of the world’s most forward-thinking AI ethicists are now calling “AI colonialism.” 

And yet, very few members of the AI pundit class seem capable of imagining, let alone understanding, AI technology (or its potential) through any other lens—while also rarely naming or acknowledging this very limitation. Tech bro fish: Meet water. You can read AI news and newsletters for weeks and months and never come across the work of decolonial AI experts Sabelo Mhlambi, Meredith Broussard, Stephanie Dick, Shakir Mohamed, Marie-Therese Png, William Isaac, or any of the folks behind Just AI, who founded a think tank responding to “the narrowness of dominant frameworks in data and AI ethics research,” or the feminist tech theorists delving into thorny issues around consent and computing. (There are exceptions to this, and they’re worthwhile.) The AI. commentariat is vastly more concerned with stuff that interests and implicates nerdy white dudes—AI porn, data security, business and finance, right-wing conspiracy theories—than about the AI industry’s exploitation of the Global South, the threat AI’s energy demands pose to the environment, or the existing exacerbation of the already dangerous and sometimes deadly racial and gender biases held by the creators and funders of AI technology.

Perhaps because the people dominating the “AI race” are themselves limited by a lessened ability to empathize with others—which is to say, they are members of the most privileged social and economic groups—something as news-making as the AI development “pause” letter can really remarkably fail to offer anything but vague hand-wringing. Signed by dozens of the most famous names in tech, the letter worries: “Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones? Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization?”

They’re not bad questions for a freshman ethics seminar—though I have trouble taking a letter signed by Elon Musk and Andrew Yang very seriously. But it’s really something to watch tech barons express awe and anxiety over jumped-up chatbots when the act of bringing actual sentient beings into the world has never been more fraught for the women and trans and genderqueer folks facing abortion bans, pregnancy criminalization, and rising maternal mortality rates since the U.S. Supreme Court struck down Roe v. Wade last summer. This is how we end up with a University of Chicago sociologist trying to make a serious argument that “we” need an “AI rights movement” when Arkansas Gov. Sarah Huckabee Sanders just signed a law that makes it easier to exploit child labor. It’s why we have (deadly!) self-driving cars but not universal health care. We have humanized computers and dehumanized people.

The risk of rushing into and through the “AI race” is not that we are about to let AI technology itself get away from us, but rather that we are in danger of foreclosing the possibility of developing decolonized artificial intelligence. If the mutually assured destruction of the AI apocalypse lies in any direction, it is in that one—the one in which we ruminate endlessly on imagined future threats from hyper-sentient robots who learned to learn from men who fancied themselves kings and captains of industry, rather than cooperators and collaborators.