Marketing7 min read

Let Me Switch It Off

Recent survey results from The Harris Poll and Quad suggest that American shoppers are less against AI than against AI they cannot control, with 73% interested in being able to switch AI features on or off, which gives marketers a practical reason to treat the off switch as a feature and not as an admission of weakness.

A blank gold switch set beside a quiet ink machine.

An off switch earns the convenience

Many brands with a shop window or an app are adding AI to the shopping journey as peak season approaches, and the temptation is to read any hesitation from customers as a lack of understanding that more exposure will cure. A survey whose AI findings Quad published on 17 November 2026 points to a different reading, which is that many shoppers see what AI can do for convenience and are interested in deciding when it does it.

The survey is “Return of Touch: The Human Advantage”, conducted by The Harris Poll for Quad, a marketing experience company, among 2,004 US adults aged 18 and over between 23 July and 12 August 2026, with a stated sampling precision of plus or minus 3.0 percentage points. Quad published its takeaways and a press release on 17 November. Sixty-eight per cent of respondents believe AI can make shopping more convenient when it is designed with their interests in mind, 54% say it helps make more time for the parts of life that matter, and in the same survey 73% said they were very or moderately interested in being able to turn AI features such as recommendations, advertising and offers on or off. The press release adds that 74% are open to AI recommendations as long as they can easily ignore or turn them off.

Through the STAR Framework, this is where the Realist lens comes in, because it values security, stability and reliability, along with simplicity and ease, and on my reading a visible off switch and plain labelling serve both halves of that description at once. The switch makes the journey feel safe because the person is never stuck with something they did not choose, and it keeps the journey simple because the choice sits where they can see it. Loss aversion, which Kahneman and Tversky described for gains and losses in decisions under risk, may offer a supporting reason, since losing control could feel like a loss that people weigh more heavily than the convenience gained, though the survey did not test that and it is my reading.

What clarity and trust look like in the numbers

The same survey shows that interest in control comes with interest in understanding. Seventy-two per cent were very or moderately interested in a clear explanation of how AI is used in their shopping experience, 81% said they trust brands more when they are clear about what is and is not AI-generated, and 75% said AI-generated content has made it harder to trust what they see when shopping online. Trust in specific uses varies, with 53% trusting brands to use AI responsibly for personalised coupons and offers, 49% for product images and videos and 48% for customer-service agents, while only 38% trust brands to use AI to change prices automatically based on demand or shopping behaviour. The report’s chart expresses the same price result the other way round, with 62% of Americans trusting brands only slightly or not at all on automatic price changes, against 45% of Gen Z and Millennials.

The Thinker lens wants the explanation, since on my reading it is the lens that asks how a feature works and what it is doing with the information it uses. Dear Algorithm describes a bank that understood this: “Monzo treated transparency as a product feature.” A clear account of where AI is used, what it decides and what it does with data is that kind of feature, and it answers the 72% directly.

The price result shows how a prevention focus may operate. Regulatory Focus Theory distinguishes a promotion focus, which is about gains and aspirations, from a prevention focus, which is about safety and avoiding loss, and on my reading automatic price changes invite a prevention response, because the worry is about being manipulated and not about missing a chance. The survey did not measure regulatory focus, so that is an interpretation and not a finding.

These figures come with several cautions, because the survey covers US adults only, it was commissioned and published by a company that describes itself as a marketing experience company, and the fieldwork took place in July and August, which means that 17 November is the date Quad published its AI findings and not the date of the answers, and the full report has been online since at least 14 October. The “interested” figures describe the share who said they were very or moderately interested in an action and not the share who would use it, and I read the report and the press release without access to the underlying data tables.

Where a new feature is a choice

The Adventurer lens, which values autonomy, exploration and freedom to act, is on my reading the one most at ease with a new feature, as long as the choice to try it is its own. A switch gives that lens something to do, since an optional, clearly labelled early version of a feature lets it go first without asking anyone else to follow.

The age figures belong here, because they are where marketers are most tempted to stop. The report says Gen Z and Millennials express greater openness than Gen X and Boomers to AI-enabled customer service and support (21% against 6%), in-store checkout (24% against 12%) and product recommendations (26% against 15%), and that their strongest interest is in practical steps such as online checkout (34%) and finding product information (29%), although it does not say exactly what those percentages measure. The Gen Z group in the survey is small, at 134 respondents, so breakdowns for it carry a wider margin of error than the overall figure.

The control and clarity figures above are reported for all respondents together, and I did not find them split by age in what I read, so I cannot say that they are equally strong in every generation. Taken at face value, though, they sit among the actions that drew the most interest on the survey’s list, alongside the 74% who were interested in a guarantee that their shopping data will not be sold, which suggests a segmentation that starts from what each customer needs from the feature, such as control, an explanation or a person, and treats age as a second cut and not the first. The segments are my suggestion, and the survey does not test them.

A person within reach

Control does not mean the absence of people. In the same survey, respondents preferred human involvement over AI in every shopping interaction measured, with 66% for customer service and support, 58% for in-store checkout and 52% for in-store navigation, and 68% were interested in reassurance that AI will assist store and service staff and not replace them.

The Socialiser lens reads a shopping experience through the people in it, so on my reading it is the lens that most wants a human available when a service question gets hard, and a brand that lets a customer reach a person without a fight is serving it. The lens readings in this article are my own interpretation of the framework and have not been tested against this survey’s data.

Here are four steps for a team preparing an AI feature for peak season, which are my suggestions and not findings of the survey:

  1. Put an off switch where the feature appears, and test whether people can find it without help, because a control that is hard to find does not do the job of a control.
  2. Explain in a sentence, in the place where the feature acts, what the AI is doing and what it does with the customer’s data, so that the question the Thinker lens asks is answered before it is raised.
  3. Offer the newest features as an optional early version, so that the people who want to go first can do so and the rest see them working before they are asked to rely on them.
  4. Keep a route to a person for service questions and say who staffs it, so that the AI is seen to help the people in the shop and not to replace them.

The survey suggests that the same feature can be welcomed for one reason by one customer and resisted for another reason by the next, and the survey’s headline results are about control, clarity and company. Whether a rollout earns the permission to be convenient depends on how many of those reasons it has answered.

If you would like to see your own decision-making style, and the lens that shapes how you judge a new technology, the STAR Cognitive Profile Assessment (33 questions) at starassess.com is a sensible place to begin.


David Chadderton is a behavioural strategist, Chief Marketing Officer and creator of the STAR Framework. He writes The Unoptimised Human.

References

The Harris Poll & Quad (2026). Return of Touch: The Human Advantage. Fieldwork 23 July to 12 August 2026, 2,004 US adults. Report file dated October 2026.

Quad (2026). What consumers think about AI and shopping: “Return of Touch” insights. 17 November 2026.

Quad (2026). 54% of Americans say AI helps make more time for the parts of life that matter. Press release, 17 November 2026.

Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52(12), 1280-1300.

Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263-291.

Chadderton, D. (2026) Dear Algorithm, It’s Not Me, It’s You: Putting the Personal Back Into Personalisation. 2nd edn. 16 May 2026.

STAR Assess

Eight minutes. Thirty-three questions.

A precise picture of how you operate: motivations, decision style, and blind spots. Free to start.

  • S
  • T
  • A
  • R
Create your account and begin