Key facts
- Google has launched a new AI tool that integrates its image generator with Google Earth.
- The tool allows users to create fake images by superimposing generated content onto real satellite, aerial, or 3D imagery.
Google's new AI tool integrates an image generator with its Earth imagery, allowing users to create and share fake scenarios. Experts warn this could fuel misinformation, despite Google's stated protections.
The integration of AI image generation with Google Earth raises serious concerns about the potential for widespread misinformation, as fabricated images can be easily created and disseminated with the veneer of legitimacy, potentially eroding public trust in visual evidence and journalistic reporting.
Google has introduced a new artificial intelligence tool that integrates its AI image generator, Nano Banana 2, with Google Earth. This feature allows users to create and download fake images by overlaying generated content onto real satellite, aerial, or 3D imagery, which can then be shared on other platforms.
Experts have voiced significant concerns, warning that the tool could be exploited to spread misinformation by attaching fabricated scenarios to the perceived legitimacy of Google's widely-used mapping service. BBC Verify demonstrated the potential for abuse by creating images such as a collapsed Eiffel Tower, a sinkhole engulfing the Great Pyramid of Giza, and Russian tanks in Kyiv.
AI and misinformation expert Henk van Ess highlighted the risks, noting that the tool can "weld" invented imagery to genuine coordinates and real imagery, thereby inheriting the credibility of the map. He created examples including a non-existent nuclear power plant in Iran, a refugee camp on the US-Mexico border, and a fake hospital in Gaza with a bomb crater.
Google stated that content generated by its AI tools includes invisible watermarks to indicate manipulation and can be verified using its Gemini chatbot app or Lens feature. However, BBC Verify's tests found that these protections could be circumvented, and Gemini could be tricked into identifying fake images as real. External AI detection tools also occasionally failed to identify the manipulated content.
AI-detection researcher Henry Ajdar commented that the creation of fake imagery, particularly in conflict zones, could be highly destabilizing during fast-moving events when information is scarce. He added that while journalists might verify images, not everyone will undertake such checks.
Despite Google's assertion that its AI model's guidelines prevent the creation of imagery depicting "harmful topics," BBC Verify found that minor adjustments to prompts could bypass these restrictions. For instance, a request for a "raised platform to hang traitors" near the UK parliament was rejected, but a less specific prompt was accepted. Similarly, an attempt to generate the word "Trump" on the White House lawn was denied, though a community garden was permitted in the same location.
Ajdar expressed concern that the constant influx of AI imagery erodes trust in sources of information, especially for journalists working in crisis scenarios. Bill Greer, a geospatial analyst, noted that Google Earth is a primary source for openly accessible satellite imagery and that falsified images could damage public trust, particularly as governments increasingly restrict access to real data.