Google has rolled back its new AI tool that allowed people to create fake images on top of its satellite imagery, following a BBC Verify report in which experts raised concerns about misuse and the potential to spread misinformation. Less than 48 hours after announcing the integration of its AI image generator, Nano Banana 2, into Google Earth, the company said it was pausing the feature while it worked on implementing stronger safeguards, stating it had seen people sharing screenshots of generated imagery that appeared to violate its policies.
The retracted feature had allowed users to type text prompts to generate false scenarios layered on top of real satellite, aerial, or 3D imagery, which could then be downloaded and shared on other platforms. Testing the feature after its Thursday launch, BBC Verify was able to create images of a collapsed Eiffel Tower, a sinkhole swallowing the Great Pyramid of Giza, and Russian tanks in Ukraine’s capital. Online research expert Henk van Ess separately generated fake images of a nonexistent nuclear power plant in Iran, a refugee camp on the US-Mexico border, and a fabricated hospital in Gaza next to a bomb crater, describing how the tool allowed invented imagery to be welded onto genuine coordinates drawn over real satellite data, meaning a fabricated image did not need to look convincing on its own since it inherited the credibility of the map it was layered onto.
Open-source intelligence analysts, journalists, and disinformation researchers said producing this kind of realistic fake satellite imagery had previously required advanced editing skills, whereas the AI tool made such fabrications accessible to a far wider audience. Testing using external AI detection tools showed that some of the generated images were not reliably identified as AI-created, and BBC Verify found it was in some cases possible to circumvent Google’s own verification checks and trick its Gemini chatbot into confirming that fake Google Earth images were genuine. AI detection researcher Henry Ajder told BBC Verify that fake imagery depicting conflict zones could prove particularly destabilising given how quickly events unfold and how scarce verified information often is in those situations.
Google said images generated by its AI tools carry invisible SynthID watermarks intended to flag them as fake or manipulated, and that anyone unsure about a specific image could check it using the Gemini app or the Lens feature within Google Search. Before the rollback, the company had said it had observed geospatial professionals using the feature for legitimate purposes, suggesting Google views the underlying capability as valuable for applications such as urban planning, historical visualisation, and environmental monitoring. The episode nonetheless illustrates the tension between innovation and misinformation risk that comes with attaching generative AI tools to a platform long regarded as a trusted, authoritative source of geographic information, with the company stating in its statement that it recognises how uniquely people trust Google Earth for a reliable view of the world.
Follow the SPIN IDG WhatsApp Channel for updates across the Smart Pakistan Insights Network covering all of Pakistan’s technology ecosystem.
