UK Trading Standards Warns AI Is Supercharging Scams Targeting Shoppers
National Trading Standards publishes its first AI scam watchlist, covering fake retailers, voice cloning, deepfake videos and romance fraud
By AXY Business Desk ·
Key points
- UK trading standards officers say AI is "super-charging" existing scams through cloned websites, fake retailers and impersonation.
- National Trading Standards has published its first AI scam watchlist, covering cloned websites, romance scams, deepfakes and personalised phishing.
- Consumer group Which? says typos or grammatical errors can no longer be relied on to spot scams.
UK trading standards officers say artificial intelligence is "super-charging" existing scams, using cloned and fake websites to trick shoppers out of their money. The warning, published on 18 October 2026, comes as National Trading Standards issues guidance for the first time on the specific AI tools and techniques criminals are using. National Trading Standards coordinates consumer protection work carried out by local trading standards teams across the country, so its guidance is intended to give those teams and the public a shared reference point.
According to the officers, criminals are building websites for fake retailers that claim to be family-owned firms selling quality UK-made products. Some of the images on these sites may show a UK High Street with Union Jack flags in shot, a visual cue designed to suggest that the business is local and established. Shoppers who order from these sites are likely to find that the product either never arrives, or that they receive a cheap, mass-produced item from a country such as China. National Trading Standards says AI-generated images, voice cloning and deepfake videos are increasingly being used to trick people out of their money.
The central message is that AI is not creating an entirely new category of fraud so much as making familiar frauds faster, cheaper and harder to detect. That distinction matters for readers, because it explains why the basic warning signs that people have learned over the years are becoming less reliable.
What the AI scam watchlist contains
For the first time, National Trading Standards has drawn up an AI scam watchlist highlighting some of the tools and techniques criminals are using. The list sets out four main areas of concern.
The first is cloned or new websites that look polished and professional in rapid time and at a low cost. A website that once took skill and money to build can now be assembled quickly, with product photographs, descriptions and reviews that look credible to a casual visitor.
The second is romance scams, in which AI is used to remember victims' personal details and keep conversations going over weeks or months. The list identifies this as a way to sustain long-term deception rather than a single attempt at fraud.
The third is AI-generated deepfake videos and voice-cloned audio used to impersonate family members or close friends asking for money in an "emergency". The emotional pressure of a supposed crisis is central to how this type of scam works.
The fourth is error-free, personalised phishing emails, text messages and social media posts, which can be generated by AI on a huge scale. Phishing is a term for messages designed to trick people into handing over passwords, bank details or payments, often by pretending to be a bank, delivery company or other trusted organisation.
How the problem developed
The scams listed in the watchlist are not new. Online shopping fraud, romance fraud and impersonation scams have been reported to consumer authorities and police for many years. Fake online shops, for example, have long been a common way for criminals to take payments without delivering goods. Romance fraud, in which a scammer builds a fake relationship online before asking for money, has similarly been a recognised problem for years.
What has changed is the cost and effort required to make these approaches convincing. In the past, a fraudster needed some technical skill, a good command of the target's language and time to write persuasive messages. Generative AI tools, which can produce text, images, audio and video from simple instructions, lower those barriers considerably. A single criminal can now produce many variations of a message, adapt them to different audiences and generate material that appears more natural than the crude scams of the past.
This is the context in which trading standards officers have described AI as "super-charging" existing scams. The phrase suggests not a new kind of crime but an increase in the scale, speed and realism of crimes that were already widespread.
Why old warning signs no longer work
For many years, consumers were taught to look for spelling mistakes, awkward grammar and unprofessional design as signs that a message or website was fake. The source material indicates that these clues are becoming unreliable. Consumer group Which? has said that typos or grammatical errors can no longer be relied on to spot scams.
This matters because many people have built their habits around these old signals. A reader who receives a clean, well-written email from a bank, delivery firm or retailer may feel reassured, when a polished message is exactly what a modern scammer would aim for. The advice to trust instinct about poor writing is therefore weaker than it once was, and people may need to check suspicious messages in other ways.
It is worth noting that the source material does not say that typos have disappeared from scams altogether. Its point is narrower: that the absence of mistakes can no longer be treated as evidence that a message is genuine.
How voice cloning and deepfakes are used
Voice cloning refers to the use of software to copy a real person's voice, often from relatively short recordings such as a social media video or a voicemail greeting. A deepfake is a realistic video or audio recording that shows a person saying or doing something they never did. Both techniques can be used to make a fraudulent request appear to come from someone the victim knows and trusts.
The watchlist highlights these methods being used to impersonate family members or close friends. A typical scenario would involve a parent receiving a call that sounds like their child, asking urgently for money to resolve a problem. The sense of an emergency is designed to stop the victim from pausing to check the story, for example by calling the family member on a number they already know.
The risk is greatest when people publish large amounts of personal material online. Photographs, videos and recordings shared on social media can provide the raw material for an impersonation. Officials and consumer groups generally advise people to think carefully about how much voice and video they make public, although this is only one part of protecting yourself, because criminals may also obtain material from other sources.
Romance scams and the long game
Romance scams stand out because they can unfold slowly. The source material notes that AI can be used to remember victims' personal details and keep conversations going over weeks or months. A scammer who recalls a person's hobbies, family news and previous conversations can appear attentive and trustworthy. Over time this can build the kind of relationship that makes a later request for money feel reasonable to the victim.
For this reason, the warning is not only about obvious fraud. Anyone who has formed an online relationship may be at risk, particularly if the relationship moves quickly, involves requests for money or gifts, or never allows a video call or in-person meeting. The danger lies in the gradual build-up of trust rather than in any single moment, which can make it harder for victims to recognise what is happening.
Protecting yourself: what the source material does and does not say
The source material states that trading standards officers have set out some ways to protect yourself, but the specific steps are not included in the verified text available for this report. Readers seeking the full list should refer to the original BBC Business article and to National Trading Standards directly.
In the absence of that list, it is useful to consider widely recognised consumer advice. Shoppers can check whether a retailer is a registered business, for example through the free UK company register run by Companies House. They can look for independent reviews rather than relying only on testimonials displayed on the retailer's own website. Anyone asked to pay for an unexpected purchase through an unusual method, such as a bank transfer to a personal account or payment in gift cards, should treat that request with suspicion.
When a message claims to come from a family member asking for urgent money, a common recommendation is to hang up and call back on a number already known to be genuine. Some families also agree a private codeword that can be used to confirm identity in an emergency. For anyone who has already sent money or shared financial details, the usual advice is to contact their bank immediately and report the incident to the appropriate fraud reporting service in their country.
These general steps are not presented here as the official guidance in the source material. They are included to help readers understand how the threats described in the watchlist are typically countered.
Why it matters to shoppers and families
The watchlist matters because the financial and emotional costs of these scams can be substantial. A fake retailer can leave a buyer out of pocket and with no goods. A romance scam can cause loss of money as well as the emotional harm of discovering that a relationship was not what it seemed. An impersonation call can pressure a family member into paying money to someone who sounds like a loved one. Because AI makes these approaches more convincing, the people most at risk may include those who have always been careful, since the usual signals of fraud are less dependable.
For consumers, the practical implication is that a message looking professional, a website appearing established or a voice sounding familiar is no longer proof that a request is genuine. Verification increasingly depends on checking through a separate channel, such as an official website typed in directly, or a known phone number, rather than relying on the content of the message itself.
What to watch next
The publication of the watchlist is significant because it gives consumers and local teams a common set of examples of how AI is being used in fraud. Analysts following consumer protection would be likely to expect that trading standards teams, banks, consumer groups and police will increasingly use this kind of guidance to explain how AI is changing scams. How quickly criminals adapt to public warnings, and whether those warnings reach the people most at risk, will be important tests of the initiative.
Readers may wish to watch whether the official advice on protective steps is published in more detail, and whether the watchlist is updated as new techniques emerge. Because AI tools are developing quickly, guidance that is specific today may need revision within a short period. The central lesson from the current warning is that the usual cues of trust, such as polished presentation, a familiar voice or a plausible story, can no longer be taken at face value without independent checking.
Questions and answers
What is the AI scam watchlist from National Trading Standards?
National Trading Standards has drawn up its first AI scam watchlist, highlighting tools and techniques criminals use, including cloned websites, romance scams, deepfake videos and voice-cloned audio, and personalised phishing messages.
Why are typos no longer a reliable sign of a scam?
Consumer group Which? says typos or grammatical errors can no longer be relied on to spot scams, because AI can generate error-free, personalised emails, text messages and social media posts on a large scale.
What do fake online retailers often promise?
According to trading standards officers, some fraudsters create websites for fake retailers that claim to be family-owned firms selling quality UK-made products, sometimes using images with Union Jack flags, while the goods either never arrive or are cheap, mass-produced items from a country such as China.
Topics: AI scams, Trading standards, National Trading Standards, Online shopping fraud, Deepfake scams, Voice cloning, Romance scams, Which?
Sources: BBC Business RSS