Fake Mould, Real Losses: AI Food Fraud Hits UK Restaurants as Brand Filings Crater

The Scam Nobody Legislated For

Picture this: a courier hands over a perfectly fine portion of jerk chicken. Forty minutes later the restaurant receives a complaint supported by a photograph of something greenish, mouldy, and apparently crawling with foreign bodies. The image was generated in seconds with a free AI tool. The delivery platform processes a refund automatically. The restaurant absorbs the cost.

This is not hypothetical. Food Alert has warned that hospitality businesses are facing a rise in AI-driven food fraud, and says it is already seeing cases in the UK, with customers using AI-generated images and AI-written complaint emails to seek refunds and compensation from restaurants, takeaways and other food operators, some images appearing to show mould, undercooked meat or foreign objects that were not present in the original meal.

More than 34 million AI-generated images are now produced every day, while AI image editing has emerged as the fastest-growing software category of 2024, expanding 441% year-on-year. What once took a Photoshop professional half a morning can now be done with a text prompt and a free account.

Two Tactics, One Blind Spot

Two primary tactics have been identified in recent UK cases. The first involves manipulated images: businesses have received complaints supported by AI-altered photos depicting alleged food safety issues such as foreign bodies or undercooked food. These images are often submitted to third-party delivery platforms, which may automatically process refunds to maintain service speed, leaving the cost to the food business even when the complaint is fraudulent.

The second tactic is the written complaint itself. A recent report by Food Alert shows how fake images of rotten or contaminated food are being used alongside AI-written complaint emails to create convincing but entirely fabricated claims. Some messages include references to regulators and legal language designed to rattle an already stretched operator into settling fast.

Food Alert said the issue of AI food fraud is intensified by how many third-party delivery platforms handle complaints, with refunds often processed automatically before costs are passed directly onto restaurants and takeaways. "There have been instances where we've seen AI manipulation of images relating to foreign bodies or undercooked food complaints," said Food Alert AI lead Alasdair Dean.

Legislation Is Playing Catch-Up

The legal gap is glaring. Current UK food safety legislation was not designed to address digitally fabricated complaints. The laws focus on ensuring safe food reaches the market, not on policing AI-generated fraud, leaving the responsibility for identifying and responding to such fraud with business operators.

UK legislation does not cover AI-generated food complaint fraud, leaving businesses exposed while regulation struggles to keep pace with the technology. One industry lawyer noted that outside sexually explicit imagery, there is currently nothing governing the generation of AI images in the UK at all. Once word spreads that the tactic works, copycats will follow.

Lawyers warn that people are "getting more sophisticated in the tools that they're using" to invent complaints, an evolution colliding head-on with the ultra-generous refund policies that helped delivery apps scale so quickly. Behind each instant refund, a kitchen absorbs a cost it cannot dispute and may not even notice.

A Brand Filings Collapse Adds Fuel

The problem lands at a particularly exposed moment for UK food businesses. According to AIBD analysis of IPO (TMD) data as of July 2026, Class 29 UK trademark filings stood at just 329 in Q3 2026, a fall of 69% against the prior period. Class 29 covers the core protein and prepared-food categories, meat, fish, dairy, preserved vegetables and edible oils, where most delivery brands compete.

The timing is painful. The UK Intellectual Property Office confirmed that official fees for trade marks, designs and patents increased from 1 April 2026, with fees rising by an average of around 25% across most services. This is the first increase to UK trade mark fees since 1998. A higher cost to register combined with a sharp drop in actual filings means that new food brands launching right now, the exact operators most reliant on delivery platforms and therefore most vulnerable to fraudulent refund claims, are going unprotected.

The result is a growing challenge for hospitality businesses already operating under tight margins and higher customer service expectations, and the lack of legislation in the UK risks even more exposure.

What Operators Can Do Now

Food Alert is urging hospitality businesses to strengthen complaint investigation procedures, maintain thorough food safety documentation as their best evidence, and work closely with delivery platforms to challenge suspicious refund claims more effectively.

Strengthening food safety management systems and maintaining thorough documentation can provide a strong defence if a fraudulent claim leads to a regulatory visit. Standardising complaint investigations ensures genuine issues are addressed and fraudulent ones are detected.

On the brand protection side, the maths still favour filing even at higher IPO rates. The UK Intellectual Property Office charges £205 to register a trade mark in the first Nice class and £60 for each additional class from 1 April 2026, with no annual maintenance fees during the 10-year protection term. For a ghost kitchen or virtual brand defending its name across three delivery platforms simultaneously, that is a modest premium for a decade of exclusivity.

The irony cuts deep. AI is simultaneously the weapon fraudsters are using to attack food businesses and the tool operators are increasingly told will save them. The difference is that the fraudsters have already deployed it.