Google removed a new AI image-generation feature from Google Earth on Friday, July 31, 2026, following concerns about fake imagery [1, 2].
The rollback highlights the growing tension between the deployment of generative AI and the need for geospatial accuracy. Because satellite imagery is often used to verify real-world events, the ability to realistically alter these images could facilitate the spread of disinformation on a global scale [2, 5].
Alphabet Inc. launched the tool to allow users to generate or modify visuals within the Google Earth platform [1, 2]. However, the company pulled the feature after it had been live for only one day [3].
Reports indicated that users were utilizing the tool to create altered satellite images that violated Google's existing policies [1, 2]. These modifications could potentially be used to manufacture evidence, or mislead the public about the physical state of specific locations [2, 5].
Google has not provided a timeline for when a revised version of the tool might return. The company's decision follows immediate backlash from experts and users who warned that geospatial disinformation could have severe implications for security and news reporting [2, 5].
The removal of the tool represents a rare and rapid reversal for the company in its race to integrate AI into its core products. By acting within 24 hours, Google attempted to mitigate the risk of its platform being used to generate believable but false geographic data [3].
“Google pulled the feature after it had been live for only one day.”
This incident underscores the unique risks associated with 'geospatial AI,' where the hallucination or intentional manipulation of imagery can be mistaken for objective physical proof. As AI tools become more integrated into mapping and satellite services, the industry must develop more robust watermarking or verification systems to prevent the erosion of trust in satellite data as a source of truth.
