The AI Sticker Tax: How to Tell Real Intelligence From Retail Theatre
Walk through a UK electricals aisle in November 2026 and you will see the same word everywhere: AI. It is on televisions, vacuum cleaners, coffee machines, mattresses, skincare gadgets and even pet feeders. Sometimes it describes something real. Often it is a sticker doing the work of an engineering department.
That matters because AI has become a premium badge. A product with the letters attached can cost more, sit higher on the shelf and sound more futuristic in the review section. This is not a guide to avoiding AI. It is a guide to avoiding paying over the odds for the word alone.
The badge is not the feature
AI-washing is the retail cousin of greenwashing: a vague, impressive claim that is hard to pin down and harder to challenge at the till. It is not usually an outright lie. More often it is a generous interpretation. A thermostat that follows a schedule becomes 'AI-powered climate intelligence'. A camera that detects a person becomes 'AI vision'. A subscription that recommends the same cat food every four weeks becomes 'personalised by AI'.
None of those things are necessarily bad. The question is whether the intelligence is doing something useful, or whether the marketing team has simply found a more exciting word for automation.
Three questions to ask before you pay extra
1. What is the AI actually deciding?
Look for a specific job. Does it recognise a stain and adjust a wash cycle? Does it detect when a room is empty and turn down the heating? Does it transcribe a meeting and separate speakers? A real feature can usually be described in one sentence without the word 'smart'.
If the answer is 'it learns your habits', ask what it learns, what it changes and how you would notice. If nobody can tell you, that is not a feature. It is a feeling.
2. What data does it need, and where does it go?
AI features often need data: audio, images, usage patterns or location. Some process it on the device. Some send it to a cloud service. Both can be fine, but you should know which one you are buying.
Check the privacy policy for plain-English answers. Does the product work without an account? Can you turn the AI feature off? Is your data used to improve a model, and can you opt out? If the answer is buried in a 40-page document, treat that as a design choice.
3. What happens when it is wrong?
Every AI feature has a failure mode. A robot vacuum may miss a cable. A transcription tool may mangle a name. A recommendation engine may keep suggesting things you do not want. The useful question is not whether it is perfect. It is whether you can override it, correct it and carry on.
Look for a manual mode, a clear undo, a human review step or a simple off switch. A product that becomes useless when the AI is switched off is not really giving you a choice.
Smart, automated, algorithmic: a quick translation
Smart usually means sensors and rules. A smart plug follows a schedule. That is useful, but it is not necessarily AI.
Automated means the product does something without you. A dishwasher has been automated for decades.
Algorithmic can mean anything from a simple formula to a machine-learning model. It sounds technical because it is vague.
AI is most convincing when it handles messy, variable inputs: speech, images, text, unusual patterns. If the task is fixed and repetitive, ordinary automation may do the job just as well for less money.
Personalised often means the service remembers your previous choices. That can be handy, but it is not always intelligence. Sometimes it is just a login and a history tab.
The premium hides in the subscription
One of the less obvious forms of AI-washing is the feature that only works with a monthly fee. The box says AI. The small print says 'AI features require an active subscription'. That is not automatically a con, but it changes the maths.
Add up the cost over three years. Compare it with a rival that does the same job without a subscription. Ask what happens if the company changes its pricing, shuts the service down or stops updating the model. A product that depends on a remote AI service is only as useful as the company behind it.
Where UK consumer rules come in
UK advertising is not a free-for-all. The Advertising Standards Authority's CAP Code requires marketing claims to be truthful and substantiated. If an advert says a product is AI-powered, the advertiser should be able to explain what that means and back it up. A vague badge that implies more capability than the product has can be misleading.
The Competition and Markets Authority also watches how businesses present technology claims, particularly when they affect consumer choice. And if a product collects personal data for its AI features, the Information Commissioner's Office's rules on fair processing apply.
You do not need to become a lawyer before buying a kettle. But you can vote with your wallet and your complaints. If a claim feels inflated, ask the retailer for evidence. If an advert seems misleading, the ASA has a straightforward complaints process. Keep a screenshot of the claim and the price you paid.
A shopper's checklist
Green flags: a named feature, a clear explanation of what it does, a manual override, an offline mode, a sensible data policy, clear update and support terms, and evidence in independent reviews rather than only on the box.
Red flags: 'AI-powered' with no further detail, a premium price justified only by the badge, core features locked behind a subscription, pressure to accept tracking, and reviews that repeat the marketing copy without testing the feature.
The test at the till
Before you pay the AI premium, try a simple thought experiment: remove the word AI from the box. Would you still buy the product at that price? If the answer is yes, the AI may be a bonus. If the answer is no, you are probably paying for the sticker.
At AIWF, we are not anti-AI. We are anti-muddle. The best AI features are the ones you notice because they save you time, not because they shout from the shelf. In 2026, the smartest shopping move is still the old-fashioned one: ask what it actually does.