Photo posted by Richard Tice raises suspicions of AI manipulation
A photo posted by Richard Tice, deputy leader of Reform UK, became a hot topic on social media after users started pointing out some pretty weird details in the image. The case, reported by The Guardian, quickly grew beyond British political debate and landed squarely in the global conversation about the use of artificial intelligence in public communication.
On Sunday, Tice shared the photo on X as proof of supporter engagement for the party in Birmingham. In the image, a diverse group of campaigners appeared under a blue sky, with big smiles and signs in hand, ready to go door-to-door for the party.
It looked like a scene that was just a little too perfect.
And maybe that is exactly the problem.
When posting the image, Tice wrote that it was the face of resilience and conviction. But it only took a closer look for dozens of X users to start questioning the authenticity of the photo. Digital intelligence experts analyzed the edited image and found a whole series of signs pointing to AI manipulation, from extra fingers and blurry faces to distorted text on signs and suspicious pixel patterns on the ground.
The party admitted the photo had been edited but insisted it was just to adjust the brightness.
Except that explanation did not convince a whole lot of people. 🤔
The episode sparked an important debate about the use of artificial intelligence in political communication and raises a question that goes well beyond the borders of the United Kingdom: where does retouching end and manipulation begin?
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
The red flags that caught everyone’s attention
When the image started gaining traction on X, formerly Twitter, it did not take long for profiles specializing in digital content verification to start dissecting every detail. And what they found was, to say the least, curious.
The most detailed analysis came from Peryton Intelligence, a digital intelligence firm specializing in hate speech and online manipulation. According to the company, the image was almost certainly generated or altered using AI. Among the most cited elements were fingers with abnormal proportions, faces slightly blurred in specific spots, and what drew the most attention: text on the signs that appeared distorted, as if it had been generated by a visual language model that still has not learned how to reproduce words coherently.
The Peryton Intelligence analysis was quite specific in its findings:
- The faces of the figures, especially around the mouth area, showed a blur effect characteristic of AI generation
- A woman wearing a denim jacket appeared with abnormally long fingers on one hand and what looked like six fingers on the other hand
- A man in a white jacket, fourth from the left, did not appear to be holding his sign at all
- The signs distorted the word Starmer in the slogan Get Starmer Out, with some appearing to display something like Get Stuppence Out
- The arrow in the Reform logo inside the letter O on the signs showed inconsistent circularity from one sign to the next
- A traffic sign in the background had a blank white box beneath it
- Houses and windows in the background showed additional inconsistencies and blurring
Anyone who has used AI image generation tools knows these are classic errors in the process, especially the difficulty with hands, fingers, and legible text. These are well-known limitations of models like Midjourney, DALL-E, and Stable Diffusion, and they remain some of the hardest things to fix without making it obvious to someone who knows what to look for.
On top of that, analysts pointed out pixel-perfect vertical lines that experts consider highly suspicious, along with irregular yet peculiarly regular pixel patterns on the ground in the scene, something that typically happens when an image has been through inpainting. This technique is used by generative AIs to fill in or replace parts of a photo. The effect is nearly imperceptible to the human eye at first glance, but when you zoom into certain areas of the image, the repetitive patterns and lack of natural texture variation give the process away. It is not magic, it is simply how these models work: they interpolate pixels based on probabilities, and that leaves traces.
What makes this case even more interesting is that the image was shared as evidence of real political mobilization. It was not a campaign graphic, it was not a promotional poster. It was, according to Richard Tice himself, a photographic record of real people supporting the party in Erdington, a suburb of Birmingham. If the experts’ findings are correct, that completely changes the meaning of what was published, turning proof of engagement into a textbook example of what we now call visual fake news. 📸
Reform UK’s response and why it fell flat
Faced with the backlash, Reform UK did not stay quiet. A party spokesperson stated that the photograph was real but acknowledged that the version published by Richard Tice had been lightly edited using AI, mainly to boost the brightness. The party pushed back on the idea that the campaigners or the photo were fake.
And to be fair, adjusting brightness and lighting would be completely normal. Taking a photo with a phone and tweaking contrast, saturation, or exposure before posting is something any content creator does without thinking twice. The problem is that brightness adjustment does not explain extra fingers, does not explain illegible text on signs, and definitely does not explain pixel patterns that only show up in images processed by generative neural networks.
The explanation sounds like a crisis management response that underestimates the technical savvy of the public. At a time when AI detection tools are increasingly accessible and popular, trying to simplify something potentially complex with a vague justification can backfire. Instead of shutting down the discussion, the party’s response only fueled the debate further, with journalists, researchers, and everyday users calling for more transparency about how the image was created and approved before publication.
Tice himself tried to redirect the narrative, saying the photograph was taken in Erdington and that there had been a dramatic shift in support for the party since 2022, when they got only 293 votes in a by-election in the area. He said the support, recognition, and energy during his recent visit were unlike anything he had seen before, and that on May 7 the Birmingham area would have strong chances of electing Reform councilors. In a general election, he said, the party could go even further and elect a member of parliament, a possibility that four years earlier seemed distant and now no longer does.
Tice’s political response, trying to highlight the party’s growth, failed to drown out the questions about the image itself. And the reaction from other parties was swift.
Zack Polanski, leader of the British Green Party, took the opportunity to take a jab: he said there is nothing real about Reform, that the party’s supposed policies for working people are fake, that the stories they tell are fake, and that now it has become clear even their campaigners are fake. A heavy statement, but one that resonated with a significant portion of British public opinion on social media.
This is not Reform UK’s first brush with AI controversy
It is worth remembering that Reform UK is a party that has been growing rapidly on the British political scene, with messaging that appeals directly to distrust of traditional institutions and mainstream media. In that context, getting caught potentially using AI manipulation to inflate a scene of popular support is especially tricky. Credibility is a political asset, and when it gets called into question over something that could have been avoided with more careful communication, the cost can be much higher than any visual benefit the image might have provided.
And this is not an isolated episode involving figures connected to the party. Matt Goodwin, a candidate who ran for Reform in the Gorton and Denton by-election, earned the nickname MattGPT after being accused of using AI to write his book. The suspicion arose because historical figures were quoted with incorrect phrases and some of the URLs in the footnotes contained the word chatGPT. Goodwin acknowledged using AI to research some data but denied that any part of the book was written by the technology.
Outside the Reform universe, another notable case involved Kate Middleton, the Princess of Wales, who was accused of using AI to edit a Mother’s Day photo published in 2024. In the image, Princess Charlotte’s left hand appeared misaligned with her coat sleeve. The case gained so much traction that the royal family issued a notice asking media outlets to stop using the photo. That episode showed that distrust of AI-edited images is not limited to partisan politics but affects any public figure who chooses to share visual content with the public.
AI in politics: retouching or manipulation?
This episode involving Richard Tice and Reform UK is part of a global trend that has been accelerating ever since AI image generation and editing tools became accessible to the general public. Today, anyone with a smartphone can use apps that apply AI to enhance, expand, or even create photographs from scratch in a matter of seconds. What used to require an experienced designer and hours of Photoshop work now happens with a single tap. And that is incredible from a technological standpoint, but it comes with responsibilities that a lot of people still have not stopped to think about.
In the political context, the line between improving a photo and manipulating a narrative can be pretty thin. An image with more light, more vivid colors, and sharper expressions might just be an aesthetic choice. But when those adjustments start altering the perception of how many people were at an event, the enthusiasm of a crowd, or even the legibility of messages the party wants to convey, the retouching stops being a design tool and becomes an instrument of deceptive persuasion. And that is where the fake news debate fully enters the conversation.
Experts in digital ethics and political communication had been warning about this scenario even before major image generation models went mainstream. The question they raise is not whether the technology will be used, because it already is, but rather what the acceptable limits are and who has the responsibility to set them. Political parties, news outlets, social media platforms, and citizens themselves all need to develop a kind of visual literacy to navigate this new environment where seeing is no longer necessarily believing. 👀
The role of platforms in detecting manipulated content
One aspect that deserves attention in this case is the role of social media platforms. X, where the image was originally posted, did not flag the content as potentially generated or altered by AI. Platforms like Meta have already implemented labeling systems for AI-generated content, but adoption is still inconsistent and automated detection tools are far from foolproof. As long as that gap exists, the responsibility for verification ends up falling almost entirely on users themselves, which is not exactly fair or sustainable in the long run.
The good news is that the online community has been getting sharper and sharper at this kind of identification. In the case of the Tice photo, it was X users themselves who first raised the suspicion, even before formal analyses were published. That suggests a growing level of digital literacy among the public, something that is essential for the health of public debate in an era of increasingly sophisticated synthetic content.
What this case tells us about the future of information
More than a localized political scandal in the United Kingdom, the episode with the Reform UK image works as a barometer of the moment we are living in. The speed with which everyday users identified possible signs of AI manipulation and brought the discussion to the center of public debate shows that awareness around this topic is growing. People are more attentive, more critical, and in many cases more technically prepared than the parties’ own communications teams. That is a positive sign in a scenario that could otherwise be pretty concerning.
On the other hand, the ease with which an edited image can circulate as absolute truth before any verification takes place still represents a real risk. The disinformation cycle tends to be asymmetric: the fake or manipulated image goes viral in minutes, while the correction or debunking takes hours or days to reach the same audience. In that gap, the narrative has already been formed in thousands of people’s minds, and undoing it is much harder than it seems.
That is why cases like Richard Tice’s matter even for people who have no direct interest in British politics. They are concrete examples of how artificial intelligence, when used without transparency, can become a tool for distorting reality with consequences that go far beyond a photo on X. And while regulations on the use of AI in political content are still in their infancy in virtually every country, the best defense available is still the same as always: question, verify, and do not share before you are sure. 🧠
