Alphabet's Google has quickly paused a new artificial intelligence image generation feature within Google Earth. The company rolled back the tool on Friday, July 31, 2026, just one day after its launch. The decision came after users created and shared photorealistic, problematic images that appeared to violate company policies and sparked widespread concerns about misinformation.[chosun+5]
AI Tool Generates False Satellite Imagery
The "create image" tool, powered by Google's Nano Banana 2 AI model, allowed users to generate custom photorealistic images. Users could zoom into any location on Google Earth's web version, tap a button, and type a prompt. The AI then layered a new, generated image over the platform's real satellite, aerial, and 3D mapping data.[wfaa+8]
The feature quickly drew criticism when users began generating convincing but false satellite imagery. Examples shared across social media included fabricated bomb craters in the Gaza Strip, non-existent nuclear facilities in Iran, and even a fictional collapse of the Eiffel Tower. Other generated images depicted widespread flooding in Washington D.C. and fires at an Iranian oil terminal, events that had not actually occurred.[chosun+5]
Open-source intelligence (OSINT) experts and disinformation researchers quickly protested. They argued that Google Earth has long served as a crucial, trusted benchmark for verifying the authenticity of conflict and disaster sites. They warned that integrating generative AI could "contaminate this standard" and "undermine public trust" in a vital resource.Henk van Ess, an OSINT expert, highlighted the danger by noting that Google Earth spent two decades building trust, only to add a button that "makes things up."[chosun+5]
Google Cites Policy Violations, Promises Stronger Safeguards
Google acknowledged the issues, stating, "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 would "roll back the feature until we implement stronger safeguards."Google did not specify the exact types of images that violated its policies.[thenationalnews+21]
Initially, Google defended the feature by explaining that all generated images were watermarked with its SynthID digital watermark. This watermark was designed to help identify content created or enhanced by AI in other Google tools like Gemini or Lens in Search.The company also noted that the AI-generated images did not appear in the main Google Earth experience for other users to see.[wfaa+15]
However, critics quickly pushed back on these reassurances. They argued that most users would not take the time to verify an image's authenticity. They also demonstrated that Google's Gemini app sometimes failed to identify AI-generated photos. Furthermore, the effectiveness of watermarks diminishes when users take screenshots, crop, compress, or repost images, stripping them of their original provenance.Experts stressed that the burden of verification should not fall on journalists and investigators.[wfaa+5]
Broader AI Challenges for Tech Giants
This rapid rollout and subsequent pause are not isolated incidents in the fast-evolving world of artificial intelligence. Tech giants are increasingly integrating AI tools across their products, but they frequently face pressure to adjust, restrict, or disable features shortly after launch due to user violations, misinformation, and privacy concerns.[ground+5]
Earlier in July 2026, Meta discontinued its "Muse Image" AI feature on Instagram just three days after its introduction. That decision followed widespread criticism over privacy concerns, including objections from a Hollywood union.Google itself faced a similar situation in February 2024 when it rolled back Gemini's ability to generate images of people after the tool produced historically inaccurate depictions.[chosun+6]
Industry experts warn that rushing AI automation without sufficient human verification and intervention could erode public trust, a core asset for technology companies.The Google Earth incident highlights how easily generative AI can undermine the trust embedded in public information infrastructure when governance fails to keep pace with innovation. It also shows a tendency for companies to prioritize commercial and consumer priorities over "reality resiliency."[chosun+1]
The company has not yet provided a timeline for when, or in what revised form, the Google Earth image generation tool might return.The incident underscores the ongoing challenge of balancing AI innovation with the critical need for responsible development and robust safeguards against misinformation.[wfaa]
