Google’s new AI satellite image tool raises fears over fake visuals

Google has introduced a new artificial intelligence feature for Google Earth that allows users to create AI generated images using satellite, aerial and three dimensional mapping data. While the company says the tool is designed to support creative and practical uses, researchers and digital investigation experts have expressed concern that it could make it easier to produce convincing fake satellite images.

The new feature, called Create Image, was launched on Thursday and uses Google’s Nano Banana 2 image generation technology. It enables users to zoom into locations on Google Earth and generate realistic visualisations within seconds using existing mapping data.

Google said the tool is intended to help people visualise historical scenes, develop real estate plans and explore other creative projects. However, experts who specialise in verifying online content warn that the technology could be exploited to spread false information and manipulate public opinion.

Open source intelligence researchers say Google Earth has long been considered one of the most trusted resources for checking locations, analysing events and verifying images. They fear that adding AI generated content to the platform could blur the line between authentic satellite imagery and fabricated visuals.

Digital investigations expert Henk Van Ess criticised the update, arguing that Google spent years building public trust in its mapping service but has now introduced a feature capable of generating fictional scenes. He warned that even if an AI generated image is not perfect, it could still appear believable because it is linked to a platform widely regarded as reliable.

Brady Africk, a research analyst at the American Enterprise Institute who specialises in satellite image analysis, also raised concerns about the technology. He said the update makes it easier to create realistic fake satellite images that could spread rapidly on social media and mislead the public.

According to Africk, the growing availability of AI generated imagery could also weaken confidence in genuine satellite photographs. He warned that journalists, researchers and fact checkers may find it increasingly difficult to distinguish authentic images from manipulated ones, making investigations more challenging.

In response to the criticism, Google said all images created through the new feature include its SynthID digital watermark technology. The invisible watermark is embedded into AI generated content to help identify that it was produced using Google’s artificial intelligence tools.

The company also said users can ask its Gemini chatbot to check whether an image contains a SynthID watermark, providing an additional method for verifying AI generated content.

The launch comes as technology companies continue expanding the use of generative artificial intelligence across their products. At the same time, governments, researchers and technology experts are calling for stronger safeguards to prevent AI tools from being used to create misleading content or spread disinformation.