A widely shared social media post claimed responders at RAF Fairford wore suspicious badges while apprehending five terror suspects, suggesting a cover-up. The post included an image made of two photos: one showing people in uniform and two men on the ground, and another cropped to focus on a badge.
AI watermark detected
OpenAI, developer of ChatGPT, has a verification tool for images and audio. It said the image was generated or manipulated using its AI tools, adding it found an OpenAI watermark.
OpenAI says detecting a watermark “indicates that the content likely originated from OpenAI tools”. It adds “detected signals are reliable, and false positives are rare”.
Comparison with media photo
Media coverage of the arrests near RAF Fairford on September 27 included a similar image, reportedly taken by a local resident. That low-quality image has blurred faces and is a wider shot, showing five shirtless people sitting on the ground, with uniformed individuals kneeling, standing, and behind them.
The images bear a strong resemblance, with similar positions and a torch-like light shining from behind. However, the AI-watermarked image is much higher resolution and has an unblurred face. The head of one uniformed person appears looking down, while the news image shows a side profile. The sleeve patch is a completely different colour in the social media image.
Arrests and bail
Five men arrested on Sunday were later released on bail. The social media post claimed the suspects had been released on bail, which matches official reports.
