Story perspectives

Google disables Earth AI image tool amid fake disasters

8/5/2026

Summary

Google Suspends AI

Google disabled the AI image-generation feature on Google Earth on August 3, 2026, a tool launched on August 2, 2026 that created artificial satellite images from prompts.
Users posted fake screenshots of a flooded U.S. Capitol, an Iranian nuclear plant and border refugee camps.
Google said some outputs breached its policies and suspended the feature while adding controls.
Researchers warned that AI-altered satellite imagery could erode trust used by journalists and investigators.

Summary

Google Watermarks Images

Google paired the Nano Banana 2 model with a SynthID watermark that tags each generated image as AI-created.
Users can zoom anywhere, click “Create Image,” type a prompt, and receive a custom scene from satellite and 3D data in seconds.
Researchers such as Henk van Ess showed the AI accepted prompts about refugees, nuclear plants and disasters, exposing gaps in content filters.

History

Historical Context

In 2018, University of Washington researchers and the New York Times released a deep-fake video of former President Barack Obama, the first high-profile demo after FakeApp made the technology public. The clip prompted studies showing most viewers failed to spot the fake, raising misinformation concerns. Today Google’s Nano Banana 2 model with SynthID watermark and warnings from Henk van Ess inform new AI-image regulations.

1 of 3

Google Suspends AI

Google disabled the AI image-generation feature on Google Earth on August 3, 2026, a tool launched on August 2, 2026 that created artificial satellite images from prompts.
Users posted fake screenshots of a flooded U.S. Capitol, an Iranian nuclear plant and border refugee camps.
Google said some outputs breached its policies and suspended the feature while adding controls.
Researchers warned that AI-altered satellite imagery could erode trust used by journalists and investigators.

1 / 3