The default whiteness of internet technologies became painfully apparent at the height of the COVID-19 pandemic. During this period, many darker-skinned students who transitioned to remote learning struggled to write exams because online-proctoring face detection tools, trained primarily on white faces, either failed to recognize their faces or erratically flagged them as cheaters when these tools did not recognize them. The racist undertones of these systems were impossible to ignore when a Google engineer had to tender a public apology after the Google Photos app mislabeled two black people as gorillas. Most recently, in 2024, Google was enmeshed in yet another bias scandal when its generative AI tool, Gemini, churned out stereotypical and ahistorical pictorial depictions of people of colour. The system was temporarily suspended because it was over-sexualizing its depictions of black women and generating ahistorical images like Asian Vikings, black female popes and black Nazi soldiers.

 Likewise, many AI voice assistants have female voices and feminine names by default. This is not by accident, as it draws directly on the social stereotype that women are naturally suited for assistant and secretarial roles. The under-representation of women in technology and the training of AI algorithms primarily on male data reflects the male domination of the technology workforce. It has meant that many technology products reinforce harmful gender stereotypes while exacerbating gender bias. In each case, the under-representation of women and people of colour in tech, along with algorithms trained on data reflecting a predominantly white, male view of the world, means these products mirror and amplify societal biases and stereotypes.

Also, shortly after Elon Musk took ownership of X (formerly Twitter) in October 2022, he revised the platform’s content‐moderation policy and reinstated accounts that had been deplatformed for spreading toxic content that sometimes helped incite public unrest. Notably, soon after, the platform’s weekly rates of hate speech climbed by roughly 50 per cent. This policy shift demonstrates that platform rules often reflect platform owners’ priorities, values and ideologies rather than serving as a neutral baseline.

Moreover, the modern computer interface, aptly named the desktop, upholds a white-collar, office-based reality. Its icons and vocabulary—including files, folders, documents, inbox and outbox—replicate a world conceived by and for middle and upper-class office workers. For users whose daily lives revolve around a mechanic’s workbench, a factory floor, or a kitchen countertop, this metaphor signals that they are entering a space structured by corporate routines. This inscription of a white-collar worldview into how we organize digital information comes as no surprise, given that these artefacts were conceived by technology designers who were raised in, orbit, and inhabit affluent, white-collar environments. These designers projected their ideological framing onto computer interface metaphors, thereby privileging their mode of work and lifestyle over alternative ones. Former Y Combinator President and tech wunderkind, Sam Altman, posited in a 2019 episode of the Conversations with Tyler podcast that:

Founders aren’t superheroes… They may play extreme sports, respond to emails within seconds, and start billion-dollar companies, but they are rarely the product of extraordinary circumstance. In fact, they tend to be solidly upper-middle class, reasonably smart, and with loving parents.

These examples make the notion of value-neutral technologies untenable. From the perfect emojis on our devices to the policies that govern what we can and cannot say online, technology is anything but neutral. Although these corporate metaphors have over time morphed into a technology lingua franca that is intuitively accessible for individuals across all class lines, this evolution only partially masks their ideological framing. Ultimately, if Silicon Valley tech barons espouse racist, sexist, white-nationalist and hyper-capitalist ideologies, then the technologies they envision, fund and build will invariably reflect, reproduce and diffuse these biases at scale the moment a product launches. However, simply diagnosing the source of bias is only half the story. What truly matters is to explore how and where these biases are wielded once a product launches.