Google Earth AI tool creates synthetic satellite images

Google has introduced an AI feature in Google Earth that generates synthetic satellite images from text prompts [1, 2].

The tool represents a significant shift in how geospatial data is presented, moving from purely observational imagery to generated content. This transition raises concerns among experts regarding the authenticity of visual evidence used in global reporting and intelligence.

Alphabet Inc. announced the feature in early 2024 [1, 2]. The tool allows users to input text descriptions to create realistic-looking aerial views within the Google Earth application and web platform [1, 2]. While the company said the tool is intended to inspire creativity, the ability to simulate geography poses a risk to public information [1, 2].

Experts said that the technology could be exploited to create false satellite imagery [1, 2]. Such synthetic images could be used to fabricate evidence of military movements, environmental disasters, or urban developments, potentially fueling misinformation campaigns on a global scale [1, 2].

The ability to generate high-fidelity imagery that mimics real-world locations makes it harder for the average user to distinguish between a captured photograph and an AI-generated simulation [1, 2]. Because satellite imagery has traditionally been viewed as a gold standard for verification, the introduction of synthetic alternatives may erode trust in digital mapping tools [1, 2].

Google has not detailed specific safeguards to prevent the generation of deceptive content, though the company said the focus remains on creative exploration [1, 2].