Google’s recent integration of an AI-powered image editing tool into Google Earth has prompted concerns among security researchers and open-source intelligence experts about the potential escalation of disinformation through fabricated satellite imagery. Announced on Thursday, the new feature uses Google’s Nano Banana image-generation model, enabling users to alter satellite images by typing desired changes and effectively “virtually reimagining” any location worldwide.
The AI tool allows users to insert or modify objects and scenes within satellite views, producing highly convincing images. Soon after the launch, users on the social media platform X shared manipulated images purporting to show a bomb crater in Gaza, nuclear power plants in Iran, and a car crash in Amsterdam—none of which reflected actual events.
Experts who rely on Google Earth as a reliable reference point for verifying imagery from conflict zones and inaccessible areas expressed alarm that the new tool could undermine trust in the platform. Sam Stockwell, senior research associate at the Alan Turing Institute, noted that Google Earth has long been used as a dated reference layer for journalists and investigators to cross-check suspicious images. He warned that embedding an AI generator into this trusted source could corrupt methods traditionally employed to expose fakes.
Echoing these concerns, Henk van Ess, an open-source intelligence specialist and author of the Digital Digging newsletter, highlighted the paradox of fabrications originating within a platform widely used for authentication. Van Ess publicly shared several edited Google Earth images to illustrate how the new functionality could complicate verification efforts.
This development arrives amid a broader climate of unease over AI-generated images, which have proliferated as tools to fabricate evidence and spread misinformation. In March, for example, satellite images allegedly altered by AI circulated on social media, claiming to depict damage to a U.S. radar installation in Qatar after an Iranian drone strike. Other posts purported to show Iran’s late supreme leader being extracted from rubble or exaggerated destruction at a U.S. military base in Iraq. These examples demonstrated how AI can be used to misrepresent real-world events and manipulate public perception.
Brady Africk, a research analyst at the American Enterprise Institute, pointed out that Google’s tool represents a significant advance beyond previous image generators because it anchors AI-generated content to authentic satellite imagery. This integration reduces many obvious artifacts typical of AI fakes, such as inconsistent terrain or scale mismatches, making the manipulations harder to detect.
Google has stated that every image created with the Nano Banana tool in Google Earth will bear a watermark detectable by its Gemini AI model. However, Africk cautioned that this safeguard might prove inadequate, as most users do not routinely check whether images they view or share are AI-generated.
As AI-generated content technology continues to evolve rapidly, the addition of such capabilities to widely used platforms like Google Earth raises complex challenges for verifying information and combating misinformation on a global scale.