Not everyone wants to use AI – but do we still have a choice?

This past July, Noella Williams was laid off from her job as social media coordinator, a role that entailed creating around 15 social posts a day. Williams, who is against the use of AI because of its environmental impact, says that she was encouraged to use ChatGPT to write the copy for these posts. Using AI wasn’t technically mandated by her employer, but the heavy workload made it hard to avoid.

“It was definitely in between the lines of encouragement and mandated,” said Williams. “I think it was encouragement, but in a way that you could only keep up with the [output] at that job if you used it.”

In the end, the ChatGPT-crafted posts never performed well anyway, according to Williams. “It felt like busy work, an overproduction of things, when, in a normal world, maybe they cut down on how much our workload was, because then it could maybe produce something of quality,” she says. “Then, when they overworked us, they're like, ‘Here’s this tool: It’s ChatGPT. Use it for this, this and this.’” 

Williams is not alone in her struggle. When ChatGPT was first released in 2022, the question of whether or not to use AI still felt like a sensible one to ask. It was, after all, an individual choice, right? Today, however, things feel murkier; many employers are convinced it’s a necessity, and tech features such as Google Search’s AI overview, which provides AI-generated summaries in response to user search queries, make it increasingly difficult to opt out of using it. With workplace pressures and automatic integrations on the rise, whether or not AI is still a choice is a question that’s become increasingly difficult to answer.

The experience of Williams has been mirrored across the workforce. Last April, for example, a leaked internal memo written by Tobi Lütke, Shopify’s CEO, went viral. Within the memo, Lutke detailed that AI would become a part of performance and peer review questionnaires, as well as the project prototyping process. Additionally, before requesting more resources, Shopify teams were asked to “demonstrate why they cannot get what they want done using AI”.

It felt like busy work, an overproduction of things, when, in a normal world, maybe they cut down on how much our workload was, because then it could maybe produce something of quality. Then, when they overworked us, they're like, ‘Here's this tool: It’s ChatGPT. Use it for this, this and this’

Employees like Anna* feel at odds with these corporate pressures. Anna works in HR, and though AI use isn’t required at her job, it’s strongly encouraged at every level, whether that means being mentioned in company-wide emails, brought up in meetings or discussed in conversation with senior leadership and management. “It’s essentially everywhere,” she says. Like Williams, Anna feels uneasy about the technology’s potential impact on the environment, noting that the increase in data centres AI will require could lead to people living in poor environmental conditions. As a 2024 study found, training an AI model can produce air pollutants equivalent to more than 10,000 round-trips by car between Los Angeles and New York City, and it’s likely that these pollutants will affect low-income communities the most. And, aside from requiring large amounts of electricity, the hardware used to create generative AI models also requires large amounts of water in order to cool, meaning that local water supplies and ecosystems may become strained.

On a more immediate level, Anna says that AI’s potential to affect critical thinking skills also gives her pause. “I don't love that, especially when media literacy is at an all-time low,” she says. “I don't think we should outsource our critical thinking skills any more than we already do.” While the science is still evolving, AI’s impact on critical thinking is already being examined. In a 2025 report from the MIT Media Lab, researchers found that participants asked to write an essay using ChatGPT had the lowest brain engagement and “consistently underperformed at neural, linguistic, and behavioral levels.” While the sample size of the study was small, the results were still jarring; English teachers described the ChatGPT-assisted essays as “soulless”.

Though companies appear keen to rapidly implement AI, the jury is still out on whether or not the technology has led to any meaningful gains for them. A 2025 report from the MIT Media Lab, for example, found that 95 per cent of organisations are seeing zero return on investment despite investing billions in these technologies. “I think that companies are probably going to lean into investing in AI for a long time because they've made that decision, touted it and put money behind it,” says Julie Carpenter, a human-AI interaction expert. “But that normalisation is changing our expectations of output and what labour looks like, and what human labour versus machine labour looks like.”

Even if you escape AI in the workplace, the way the technology is becoming integrated into platforms we use every day still makes it impossible to ignore. Microsoft, for example, recently bundled Copilot AI into its 365 package, adding the AI service to Word, Excel and other apps within the 365 suite. Going back further, predictive text, that is, AI-generated text that writes responses like “thank you!” or “sounds great!” for you, has existed in Gmail since 2018. Integrations like these, which Carpenter calls “an interesting form of consent collapse”, were the subject of a July Substack post by tech reporter Brian Merchant.  

“Silicon Valley’s preferred narrative is that the generative AI boom is what the people want; that it’s been driven by consumer demand and organic user interest,” writes Merchant. “But the details of how AI features have been developed, deployed and served to consumers on the biggest pieces of digital infrastructure paint a starkly different picture.” 

Merchant’s article cited research by design scholars Nolwenn Maudet, Anaëlle Beignon and Thomas Thibault, whose 2025 study, "Imposing AI: Deceptive design patterns against sustainability", explored the ways companies such as Google, Adobe, and Snapchat have deployed AI. The study called out countless ways AI features have been designed to be unavoidable; among them, the fact that AI features are given the most valuable space in interfaces, are easy to trigger but difficult to deactivate, and are given more dedicated buttons to encourage interaction.

“It really signals that they're trying to force people to use the tool in a way that they probably wouldn't have if they were doing it of their own accord,” Alex Hanna, a sociologist and Director of Research at the Distributed AI Research Institute. Hannah also noted that AI integrations often use “dark patterns”, a UX design strategy that intentionally tricks the user into doing something they did not intend to do. 

Silicon Valley’s preferred narrative is that the generative AI boom is what the people want; that it’s been driven by consumer demand and organic user interest. But the details of how AI features have been developed, deployed and served to consumers on the biggest pieces of digital infrastructure paint a starkly different picture

So, is AI use still a choice? Pace University’s James Brusseau, a philosophy professor whose current research explores humans’ experience with artificial intelligence, thinks that we may be beyond that point. Still, even he, someone who believes in using AI, recognises its drawbacks. 

Brusseau points out that when he was a student, researching something online meant more than relying on an AI overview. “Let me give you an example,” said Brusseau. “Back when I was your age, and I was doing a Google search, I would get that list just of links. Half the time, my answer would be in the first one or two links,” said the professor. The other half of the time, the professor recalled, he’d go down the search results line by line, and find “a different way of thinking” about his idea. 

“We don't know how this is gonna turn out. We don't know if in the end, the efficiency gains will be worth the loss in serendipity, worth the loss in new discoveries,” he says. As Brusseau points out, only time will tell how the effects of AI will play out. But the professor argues that nuance is key, and that it’s important to ask the right questions in the first place. 

“I don’t think it’s ‘use AI’ or ‘don’t use AI’,” says Brusseau. “It’s about how you use AI. And this tension between that – it’s like a drug. The narcotic of convenience.”

As Brusseau notes, there’s a simplicity to using AI. But what do we lose when we succumb to that “narcotic of convenience?” While AI is supposed to make our workloads easier, it now seems that actually avoiding AI has become work in and of itself; if you don’t want to use it, that’s an active and intentional choice that’s becoming harder and harder to make. Whatever friction that once existed is now dwindling.

“Freedom and creativity require work,” says Brusseau. “And convenience is always seductive.”

* Name has been changed to protect anonymity