Google U-turns on AI Earth image feature after users ‘collapse Eiffel Tower’

The tool could create ‘visual stories’, Google said (Picture: Google) A new AI tool that lets users make phoney edits of real Google Earth satellite images has been pulled only 24 hours after launch. The tech giant rolled out a free tool worldwide on Thursday that lets users create ‘visual stories’ of real places using its powerful image generator. Showing off the feature, Google suggested city slickers could turn a car park into a garden or a teacher could show what a site looked like in 78 AD. Yet experts and social media users instead found that it could edit a ‘plane hitting a Manhattan skyscraper’ or a fake nuclear plant in Iran. AI researcher Henk van Ess wrote in a blog post on Thursday: ‘Tonight I typed just one sentence into Google Earth and put refugees near the Mexican border. Then I planted a nuclear plant in Iran. ‘Then I put a fatal crash on a street in Amsterdam. Google’s own satellite imagery underneath all three. ‘What on earth is Google doing?’ What was the feature? Google Earth is a spatial information service made of a patchwork of real satellite, aerial and 3D imagery captured by Google over 20 years. By clicking on a ‘Create Image’ tab on the web version of the app, users could type a prompt – an instruction for an AI – to ask it to edit a site. Google’s examples show the AI-generated image isn’t always photo-realistic and could be stylised like a cartoon. Not all the AI-made images were photorealistic, as this demonstration from Google showed (Picture: Google) The original Google Earth frame could also be visually different, such as turning a bright blue sky into a sunset. In this example from Google, the user makes a mock of their dream house. Not only is the house added, but the scenery and sky change (Picture: Google) Technology news outlet 404 Media’s tests found the tool could add a ‘bomb blast and crater’ to Los Angeles within 10 or so seconds. The BBC, meanwhile, forged images of a ‘collapsed Eiffel Tower, a sinkhole swallowing the Great Pyramid of Giza and Russian tanks’ in Kyiv, Ukraine. ‘You don’t get better fakes, you get far more of them’ Van Ess told Metro that making fake images is nothing new – but the ease of making them is. ‘Every fake you could make with this tool was possible in Photoshop 20 years ago,’ he said. ‘The barrier was never skill, it was effort.’ Van Ess recorded himself prompting Nano Banana, Google’s AI model for image generation, to insert a ‘traffic accident’ in Amsterdam. ‘And crowd,’ the prompt added, ‘make whole picture in high-resolution.’ The model forged an image of a red car that had smashed into the side of ROBO-BAR in De Weteringschans, with shattered glass covering the road. The refugees I planted tonight are still sitting on that road. The nuclear plant is still in Iran. They only need to escape this page once, and some of you would believe them. (5/5) pic.twitter.com/07mrHmCU73— 𝚑𝚎𝚗𝚔 𝚟𝚊𝚗 𝚎𝚜𝚜 (@henkvaness) July 30, 2026 Van Ess said that Google Earth has long been used by digital sleuths and journalists who examine open-source intelligence, or OSINT. ‘A satellite image isn’t a memory, it’s a record – courts, insurers and war crimes investigators treat it as evidence of what stood on a patch of ground on a given day,’ van Ess said, pointing to warzone reporting. That’s why adding a tool that uses the same technology that ‘removes tourists from holiday photos’ might not be a ‘good thing’. ‘You don’t get better fakes, you get far more of them, from people who’d never have bothered,’ van Ess added, ‘And they arrive welded to real coordinates, in the same light and at the same angle as the genuine capture beside them. ‘The forgery doesn’t have to be convincing on its own. It inherits the credibility of the map it came from.’ Google pulls AI Earth tool until it strengthens ‘safeguards’ Google said on Friday that it takes ‘misinformation seriously’. All content made by Google AI content tools has an invisible watermark, called SynthID, that Google’s chatbot Gemini can spot. In a post on X, Google added: ‘We know that people uniquely trust Google Earth for a reliable view of the world. ‘We’ve seen geospatial professionals using this feature for a range of useful purposes, however, we’ve also seen people sharing screenshots of generated imagery that appear to violate our policies. Re: our statement on Image Generation in Google Earth: “We know that people uniquely trust Google Earth for a reliable view of the world. We’ve seen geospatial professionals using this feature for a range of useful purposes, however we’ve also seen people sharing screenshots of…— News from Google (@NewsFromGoogle) July 31, 2026 ‘So we’re rolling back this feature in Google Earth while we work on implementing stronger guardrails. ‘It’s important to note that generated images didn’t appear in the main Google Earth experience for others to see and were watermarked as AI-generated.’ Google did not specify which images violated its policies but said they were not part of the main Google Earth experience and were watermarked. Van Ness questioned, however, whether Google’s safeguards can ever hold up, especially as AI images degrade every time they’re screenshotted. ‘So keep AI on the reading side of satellite imagery,’ he added. ‘Shelve it on the painting side.’ Meta was forced into a near-instant 24-hour U-turn last week when it unveiled an AI tool that could edit people’s public Instagram posts. Users had to manually opt out if they did not want their public photos to be available to be remixed by the social media titan’s model, Meta AI. The company admitted three days later that it ‘missed the mark’ with the feature. Get in touch with our news team by emailing us at webnews@metro.co.uk. For more stories like this, check our news page. Arrow MORE: Schools warned about unknown dangers of learning devices taking over classrooms Arrow MORE: OpenAI’s rogue robot tried to hack into other companies – could your data be next? Arrow MORE: Your chat with the AI chatbot Claude could be publicly available online News Updates Stay on top of the headlines with daily email updates.