An AI-generated image purporting to show Thai police officers in elaborate festival dresses flanking a handcuffed drug suspect was so convincing that it landed on the front page of Britain’s Daily Star and was published by The Telegraph, The Sun and the New York Post — before anyone realised the picture was a complete fabrication.
The photograph, released on the Facebook page of Tha Luang police station in central Thailand, depicted five men and one woman in skin-tight sequins and feathers, smiling around a suspect they had apparently caught during an undercover operation. The Sun described the officers as a “burly crew” who “slipped into skin tight sequins and feathers for the covert mission in Thailand”. The Daily Star reported that “the team of five blokes and one woman shared a snap of themselves in frilly dresses with the nicked suspect on Facebook”.
The story was irresistible — and entirely misleading. The arrest itself was real, but the image was generated by artificial intelligence. The genuine photograph, later posted by the same police station, shows the five male officers in their regular uniforms. The woman dressed as a dancer was never there. The image was a fake from start to finish.
The Truth Behind the Picture
Police Sergeant Rachata Mitrsuripong, the administrator of the Tha Luang Provincial Police Station’s Facebook page, admitted creating the AI image. He said his aim was to project “a friendlier image” of the force and to show “a cute and humorous side”, explaining that he wanted the public to feel more comfortable approaching officers. The station’s superintendent, Panthep Panadit, confirmed that the image had been produced using AI software and said he had not posted it himself, having been sent it for review.
The picture was picked up first by local Thai media before spreading rapidly around the world. Among the outlets that ran it were the Daily Star (which put it on its front page), The Telegraph, The Sun, the Daily Mail, the Mirror, GB News, the Express, the New York Post and CNN Brazil. Several, including The Sun and the Daily Mail, later updated their stories or published new ones clarifying that the image was AI-generated. The Tha Luang police station also amended its Facebook post, adding the original photograph and a note that the previous one was not real.
The Verification Crisis for Newsrooms
The episode has laid bare a growing crisis for media organisations: how to verify images when even seemingly official sources cannot be trusted. AFP fact-checkers identified the fake using Google’s SynthID detection tool, which revealed a star-shaped Gemini digital watermark and gave a “very high” degree of confidence that the image was created with Google’s AI tools. But such technology is not considered reliable enough for routine use, and there is no foolproof method to confirm an image’s authenticity without a direct relationship with the photographer.
Newsroom editors describe the task of vetting images as increasingly time-consuming and precarious. The problem is compounded by the creeping use of AI-generated imagery inside official channels — in this case, a police department’s own social media account. As a result, editors are braced for the reality that it is unlikely every AI image will be spotted before publication.
The industry is responding in a variety of ways. Some news organisations now add disclaimers when sharing third-party images, stating that they have verified the validity of the footage or photograph. However, the effectiveness of such labels is debated, with many journalists arguing that it is more important to explain how verification was conducted. Other projects are exploring blockchain technology as a means of authenticating images in journalism.
Complicating matters further, the same technology has created an opposite challenge: viewers increasingly suspect genuine images of being AI-generated, forcing editors to defend real photographs from unfounded claims of fakery. The dual pressure — spotting fake images while proving real ones are authentic — places an extraordinary burden on newsrooms.
This was not the first time Thai police have used AI in ways that confused the public. In late 2025, Thai border police apologised for posting an AI-manipulated image that falsely depicted flood relief rescuers carrying automatic weapons and combat gear. That image was also identified using Google’s SynthID tool, and the stated reason for the manipulation was to demonstrate readiness, despite a government warning against fake news. In January 2026, officers from Samchuk Police Station in Suphan Buri used AI to digitally edit themselves into Disney princess and anime (One Piece) outfits for social media posts, with the aim of making police communication more engaging and protecting officers’ identities during actual operations.
Beyond these public-relations stunts, Thailand has seen a sharp rise in AI-powered crime. In May 2026, six Nigerian nationals were arrested for allegedly operating a deepfake romance scam syndicate that targeted older Thai women using fake online identities and manipulated video calls. In a separate case, 11 suspects were arrested in connection with a scam operation linked to a weapons cache. The Royal Thai Police have warned the public about the surge in deepfake AI scams, which can impersonate trusted individuals or organisations.
The Tha Luang incident underscores a fundamental shift in the media landscape: the traditional notion that “seeing is believing” no longer holds. The erosion of trust — both in the images that accompany news and in the official sources that supply them — is a challenge that editors acknowledge cannot be solved overnight. With each AI-generated hoax that slips through, the credibility of journalism itself is eroded, and the task of restoring it grows only more difficult.
