The growing tension between AI use and AI avoidance
More people are trying to cut back on artificial intelligence. But a new report suggests that stepping away from AI entirely may already be off the table.
A survey of 2,055 UK adults found that 42% deliberately limit how much they use AI. Yet 70% of respondents said avoiding exposure to AI would be difficult or impossible — even when they actively wanted less of it. The data paints a stark picture: rejecting a chatbot is easy. Escaping the automated systems embedded in workplace tools, banking apps, or streaming services is not.
Deleting an app like ChatGPT only removes the most obvious form of AI. People have far less control when those systems sit inside software they already depend on for work or daily life.
Why people are setting boundaries
Privacy is the single clearest reason people are drawing lines. Twenty-nine percent of respondents cited concerns about data privacy, security, and compliance — making it the most common motive for limiting AI use.
Another 22% said they simply prefer to keep working the way they already do. Notably, lack of skill wasn’t the main obstacle. Many people understand the technology and still decide it doesn’t belong in their routine. That cuts against the industry’s favorite assumption: that resistance fades once people become familiar with AI. Some users know exactly what these tools offer and still don’t think the tradeoff is worth it.
The findings echo broader worries about AI privacy concerns in consumer tech, where data collection often runs in the background without clear user consent.
Can opting out still count as a real choice?
Consent gets weaker when people can’t tell where AI is operating or what information it’s processing. Choosing not to open ChatGPT is straightforward. Avoiding automated decisions inside another service — like a hiring platform or a credit scoring tool — is much harder.
People can reject the visible tools while hidden systems keep running in the background. A meaningful opt-out would require clear disclosure, usable controls, and an alternative that doesn’t punish anyone for declining. Without those choices, AI exposure becomes the default. Consent then starts looking less like a decision and more like a buried settings menu.
This dynamic is especially troubling in workplace environments, where employees may have no say over the hidden AI systems embedded in productivity software or HR platforms.
What a real opt-out would need
- Clear, upfront disclosure of where and how AI is used
- Simple controls that don’t require digging through menus
- A non-AI alternative that doesn’t penalize the user
- Guarantees that opting out won’t degrade service quality
Why the backlash may keep growing
Public sentiment is already shifting. The share of UK adults who believe AI carries more risks than benefits rose from 48% in 2023 to 52% in 2026. Those seeing more benefits fell from 38% to 34%. That’s a significant swing in just three years.
Gen Z makes the picture even messier. Younger adults use AI more often than older generations, but they’re also more likely to restrict it and worry about the risks. They’re the heaviest users — and the most skeptical.
That paradox matters. It suggests that familiarity with AI doesn’t necessarily breed acceptance. In some cases, it breeds caution. The more people understand how these systems work, the more they may question whether the tradeoffs are worth it.
Employers and tech companies now have a blunt warning from this data. Clear disclosure and practical opt-outs may decide whether caution stays manageable or turns into wider rejection. The window to build trust is narrowing.
For now, the message from users is consistent: they want control. And they’re increasingly aware that they don’t have it.
