My frustrations with AI over the last year have grown exponentially. Not only because of its ubiquitous presence, but because its integration has felt out of my control. Nearly every major platform on the internet now has a chat box, personal agent, or ‘friend’. Not only did I not ask for this in the first place - the ads, the constant pressure to adapt or fall behind - but now I can’t even opt out of it. And with growing reports of AI deteriorating our critical thinking skills, the environmental impact data of centres, and concern for its effect on the job market, it’s a tough pill to swallow. 

Before AI models were woven into our everyday lives, their creators were already disregarding consent. The list of public facing trials against AI companies is extensive, mostly for copyright infringement through their illegal usage of media information to train their models. For example, in 2023 the New York Times sued OpenAI for scraping its website for “millions” of articles to use in their model training process. In these cases, it’s not only the media companies which are affected, but also their contributors, whose writing and ideas have now been stolen without their knowledge and permission. 

“Consent has to be specific, informed, freely given, and revocable,” explains Meg Leta Jones, Professor of Communication, Culture, and Technology at Georgetown University. “Tech consent flows usually fail at least a couple of those.” Some positive legislative changes have led to greater user autonomy, mainly with just-in-time pop-ups, where an app will ask you a specific request at the exact moment you try to use it. Leta Jones explains that Apple does this with their “App Tracking Transparency”, where they ask you if they can continue allowing an app to track other apps and websites, or when you get prompted to allow access to your camera or location.

These consensual checkpoints are only effective in particular circumstances. Think about how many times you are prompted for your consent for cookie collection, or the 50-page privacy notice you get every time your phone updates. At this point, you probably just click through the carousel of questions, which is exactly the problem. These pop-ups appear to give you some agency, but the choice is illusory: realistically, no-one is going to read through fifty pages of terms and conditions just to download an app, nor would everyone have the legal knowledge to truly understand what they were agreeing to if they did. Saying no to cookies can sometimes come with a monetary cost. Saying no to a privacy notice might mean you can’t update your device. The consequences here acts as a powerful incentive – the refusal to consent often means no updates, the device slowing down and certain applications no longer working, until eventually you’re left with a barely functioning black box. So sure, it’s a ‘choice’, but clearly a coercive one. 

We also have the choice to reject technology altogether - deleting our Tiktok accounts, chucking our smartphone into the Thames - but this can carry serious social implications. “Platforms have increasingly become the infrastructure of our social, economic, and civic lives,” explains Helen Hester, Professor of Gender, Technology, and Cultural Politics at The University of West London. “Refusal is then pretty costly, which makes exiting impractical. You cannot meaningfully consent when refusing means losing access to social life itself – to aspects of work, communication, political participation, and basic coordination.” 

We also can’t opt out of other people’s behaviour. Wearable filming accessories, like Meta Ray-Ban and Google Warby Parker glasses, create another unique consent problem. There are no laws against filming people in public, but there are still some social norms which make it awkward: if someone catches you pointing your phone at them and snickering, they’re probably going to clap back (and rightly so!). The problem with the new slew of wearable tech products is that they’re so discreet, meaning the people being filmed are often unaware. In an effort to curb this problem, newer AI glasses now have a built-in light that flashes when the user is filming, but lots of people will have no idea what this means, and even those that do could easily fail to notice. Besides, it’s hardly fair to expect us to be quite so vigilant every time we leave the house.

“What does consent mean when you are not the user of a system, but are nonetheless continuously processed by it?” Hester posits. This sort of ambiguity opens up the possibilities of harm. “The risk is not only overt misuse, such as covert sexualised recording or upskirting-type practices. It is the broader normalisation of ambient scrutiny, where being visually present increasingly means being legible to machine systems you did not consent to enter.” This can already be seen with the popularity of Instagram accounts dedicated to men hitting on women while using these sorts of glasses. If being harassed by men in public isn’t bad enough, now it can be filmed from their POV and posted for millions to watch online (it’s also worth noting that these practices disproportionately affect surveillance-vulnerable groups, such as sex workers).

“It is not simply that consent mechanisms are failing, it’s that ‘consent’ is being asked to do work it cannot do,” says Hester. The limited options we are being served today are not informative enough for the user to fully comprehend what they are agreeing to. True consent entails understanding, alternatives, and freedom from coercion, but the reality is no one has time for that. “I don’t particularly want to exert myself figuring out the ins and outs of the contemporary legal platform landscape – and I research digital culture for a living!” Hester quips.

So, are we in a tech consent crisis? Hester says yes, undeniably, but with the caveat that making ‘digital consent’ more meaningful and robust is, at this point, not enough to address the way Big Tech continues to encroach on our freedom and privacy.  Instead, we should be demanding more public governing and collective agreement over how these technologies are implemented into our lives, which includes questioning whether they are really necessary or desirable to begin with. If we want to opt out, we cannot do it alone– we need a full factory reset.