Google rolls back AI image generation in Google Earth over policy violations

Alphabet's Google has decided to roll back an AI image-generation feature introduced in Google Earth just a day prior, following user-generated images that seemingly violated the company's policies. The tool, powered by Google's Nano Banana 2 AI model, enabled users to create photorealistic images based on Google Earth's extensive satellite, aerial, and 3D imagery by simply typing a text prompt for any global location.

Concerns were raised about the feature's potential for spreading misinformation, as it allowed for the creation of fabricated scenarios overlaid on real imagery of landmarks and monuments. Google acknowledged the issue, stating on X that while geospatial professionals found the feature useful, "people sharing screenshots of generated imagery that appear to violate our policies" prompted the decision.

“

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.

The company announced it was pausing the capability to implement stronger safeguards. Google did not specify the exact nature of the policy-violating images but confirmed they did not appear in the main Google Earth interface and were watermarked as AI-generated.

This incident underscores the broader trend of major tech companies rushing to integrate AI across their products while simultaneously needing to adjust, restrict, or temporarily disable features due to user violations, misinformation concerns, and policy issues. Meta, for instance, recently discontinued its Muse Image AI feature after facing widespread criticism over privacy concerns.

“

the images did not appear in the main Google Earth experience and were watermarked as AI generated.