AI Halloween Scams UK 2026: How to Stay Safe

Halloween has grown into one of the UK's biggest shopping weekends, and fraud researchers say it has grown into one of the biggest scam weekends too. Security firm SecurityBrief UK found that 63% of Halloween-themed spam and adverts scanned in recent research were designed to steal money or login details, using fake giveaways, cloned retailer branding, and malicious social media ads.

What has changed since last Halloween is the tools behind these scams. The same AI techniques already causing problems at Christmas, generating convincing fake shopfronts, writing spam that slips past filters, and producing adverts realistic enough to fool most people, work just as well applied to a costume search in October. Our guide to AI Christmas scams covers the festive version of this pattern in detail; this guide covers the Halloween one.

This article covers the AI-powered Halloween scams most likely to reach UK shoppers in 2026, how to spot them, and exactly what to do if you are targeted. For the basics on staying secure with AI all year round, see our guide to staying safe with AI.

Why AI makes Halloween scams harder to spot than ever

AI has made two things much easier for scammers: building a website that looks completely genuine, and writing a message that reads naturally enough to avoid raising suspicion. Neither used to be true. A fake shop once looked amateurish. A scam email once had spelling mistakes or clumsy phrasing that gave it away.

Researchers at Check Point found more than 10,000 fake retail adverts being created daily on social media in the run-up to the 2025 Christmas season, many using AI-generated product photography that is nearly impossible to tell apart from a genuine listing. There is no reason to expect Halloween shopping is treated any differently by the same fraud networks, since it draws the same volume of searches for costumes, decorations, and party supplies in a short window each year.

The clearest sign of how convincing this has become came from Santander, which ran a public test earlier this year showing UK adults ten intentionally fake AI-generated Christmas adverts. Only 26% could correctly identify which ones were fake, meaning three out of four people could not. The bank's own press release on the campaign makes the point plainly: AI-built adverts and fake shopfronts have crossed a line most people cannot see past unaided. There is nothing about that capability that switches off between one seasonal shopping period and the next.

Five AI-powered Halloween scams to watch for

1. Fake costume and decoration shops built with AI

You search for a costume or a specific decoration and find a website offering it well below the price anywhere else. The photos look real. There is a returns policy, a checkout page, and sometimes reviews. You pay, and either nothing arrives or a cheap, unrelated item turns up weeks later.

Research into Halloween fraud by SecurityBrief UK found "counterfeit storefronts" cloning real, recognisable retail branding to direct shoppers to fraudulent checkout pages designed purely to harvest payment details. What makes this harder to catch in 2026 is that the product photography no longer has to be stolen from the real retailer. AI image generators can now produce a costume photo that never existed, styled exactly like a professional product shot, in seconds and at almost no cost.

Before buying from an unfamiliar site, search the shop's name alongside the word "reviews" in a separate browser tab, away from any link you were sent. A genuine UK retailer will have a verifiable Companies House registration and reviews across more than one independent platform. Pay by credit card where you can. Under Section 75 of the Consumer Credit Act, your card provider shares liability with the retailer for purchases over £100 if the goods never arrive.

2. AI chatbots recommending scam Halloween shopping sites

If you ask an AI assistant such as ChatGPT to help you find a good deal on a costume or decorations, treat any unfamiliar site it names exactly as you would treat one from a search engine or an advert: with a healthy amount of doubt until you have checked it separately.

National Trading Standards has warned that criminals are deliberately manipulating AI models into recommending fraudulent websites as if they were genuine, trusted retailers, a tactic sometimes called AI poisoning. The organisation has already documented cloned versions of well-known UK retailers appearing inside AI shopping answers, complete with implausibly large discounts. Louise Baxter, head of scams at National Trading Standards, has made clear that an AI recommendation on its own is not proof that a website is genuine. Full detail is in the National Trading Standards warning, and the mechanism applies equally whenever you use an AI tool to shop, not only at Christmas.

An AI recommendation is a starting point, not verification. Check the retailer independently before entering any payment details, the same way you would for a site found any other way.

3. AI-written "free Halloween treats" giveaway scams

An email or text arrives offering a free bundle of sweets, a discount code, or a prize draw entry tied to a well-known brand, timed to land in the days before Halloween. The writing is polished and professional, with none of the odd phrasing or spelling slips that used to be a reliable warning sign.

This pattern shows up at scale in SecurityBrief UK's analysis: seasonal giveaway emails impersonating major retailers, some using an unusual filter-evasion trick of embedding unrelated text, such as references to photosynthesis, to slip past spam detection. Researchers described the technique as "rarely seen outside more sophisticated phishingA scam that tricks people into giving personal information. operations." AI writing tools make this kind of message quick and cheap to produce in bulk, easy to personalise so it reads as though it was written for you specifically.

Genuine retailers do not usually run surprise giveaways delivered by unsolicited email or text with a link to claim them. If you want to check whether an offer is real, go to the retailer's official website directly by typing the address yourself, rather than clicking through from the message.

4. Malicious social media adverts for Halloween sales

A sponsored advert appears in your social media feed offering discounted Halloween decorations, costumes, or party supplies from a brand or shop you half-recognise. Clicking through can lead to a fake checkout page designed to steal card details, or in some cases to a page that installs malware disguised as a legitimate app or tool.

One specific pattern flagged by SecurityBrief UK's research: sponsored adverts on major social platforms redirecting users to malware-hosting pages disguised as ordinary shopping or trading tools. Social media platforms do check adverts before they go live, but AI-generated creative and rapidly rotating fake accounts make it easier for fraudulent adverts to slip through before they are reported and removed.

Treat a sponsored advert with the same caution as an unsolicited email. Do not tap through and enter payment details on a page you reached from an advert without first checking the seller independently.

5. Deepfake video adverts, and voice-cloned Halloween calls

A video advert or social media clip appears to show a familiar face, sometimes a celebrity, sometimes what looks like an ordinary customer review, enthusiastically recommending a Halloween deal or product. The voice and face both look convincing.

Santander's own test, referenced above, demonstrated exactly how far AI-generated video and imagery has come using nothing more than off-the-shelf tools available to anyone. The same technique used to build those ten fake Christmas adverts works just as well for a Halloween sale video, and there is no technical reason it would be limited to one shopping season. Treat any video endorsement of a product or shop with the same scepticism you would apply to an unfamiliar written advert: a convincing face and voice is no longer proof that a deal, or the person recommending it, is genuine.

The same cloning technology also shows up by phone. A call claims to be your bank, warning of "suspicious activity" on a card just used for Halloween shopping, and asks you to move money to a "safe account" to protect it. The voice can now be cloned from only a few seconds of audio, so it may sound exactly like a real member of staff. No genuine bank will ever ask you to transfer money to a new account to keep it safe. Hang up and call your bank back on the number printed on your card, not one given to you during the call. Our guide to AI voice scams covers how this cloning works and how to set up a family safe phrase.

Warning signs across all Halloween scams

These patterns show up across nearly every version of Halloween fraud, whatever form it arrives in.

• Prices far below anywhere else. A costume or decoration priced well under every other retailer is a common lure, not a genuine bargain.
• Unexpected contact. A giveaway, discount code, or prize you did not enter for or request deserves scepticism, however official it looks.
• Pressure to act fast. "Today only" and "while stocks last" wording exists to stop you pausing to check.
• Unusual payment requests. Bank transfers, gift cards, or cryptocurrency offer no buyer protection if something goes wrong.
• No verifiable trading history. A shop with no Companies House registration and no reviews outside its own site is a strong warning sign.
• It resists a quick, independent check. A genuine seller's details hold up when you search for them separately. A fraudulent one usually does not.

What to do if you are targeted

You are always allowed to stop and check before you pay or share anything.

1. Do not click links in unexpected texts or emails. Go to the retailer's official website by typing the address yourself.
2. Search the shop or offer name separately, alongside "reviews" or "scam". Do this in a fresh browser tab, not by following a link from the message.
3. Pay by credit card where you can. Section 75 of the Consumer Credit Act gives you a claim against your card provider for purchases over £100 if goods never arrive.
4. Tell someone. A family member or friend, especially before you pay for anything from a seller you have not used before.
5. Contact your bank immediately if you have already paid or shared card details. Dial 159 to be connected directly to your bank's fraud team. This works for most major UK banks, including Barclays, Halifax, HSBC, Lloyds, NatWest, and Santander.

Under rules introduced in 2024, UK banks are required to reimburse most victims of authorised push payment fraud, where you were tricked into transferring money yourself. Report to your bank promptly, then to Report Fraud, since acting quickly generally means more options remain open to you.

If building more confidence with AI tools, including how to spot when something is not what it claims to be, is something you would find useful, WellWired Academy is a short online course built for UK adults who are new to AI.

How to report a Halloween scam in the UK

Report Fraud (formerly Action Fraud) is the UK's national fraud reporting service. Report online at reportfraud.police.uk or by calling 0300 123 2040 (Monday to Friday, 8am to 8pm). In Scotland, report to Police Scotland on 101.

Suspicious texts: forward the message to 7726, free on most UK networks and spells SPAM on a standard keypad. Your mobile provider will investigate and can block the sender.

Suspicious emails: forward to report@phishing.gov.uk. The National Cyber Security Centre reviews these reports and works to take fraudulent sites down.

Reporting matters even when you were not caught out and lost nothing. Every report helps build the picture that lets police and platforms act faster against the same fraud reaching someone else. The Take Five to Stop Fraud campaign sums up the approach in three words: Stop, Challenge, Protect.

For scams that reach you by phone specifically, our guide to AI phone scams in the UK covers voice cloningAI that can mimic a person's voice using samples. and other call-based fraud in more depth. For general AI safety across the rest of the year, see our guide to AI voice scams.

FAQ