Google Withdraws New Earth AI Tool After Warnings Over Misinformation Risks

Tech Giant Suspends AI Satellite Image Feature Less Than 24 Hours After Launch Following Deepfake Concerns                         

                                       

Google has temporarily withdrawn a newly launched AI-powered image generation feature for Google Earth after widespread criticism from researchers, journalists and open-source intelligence (OSINT) experts who warned it could be exploited to create highly convincing fake satellite imagery and spread misinformation.

The feature, powered by Google's Nano Banana 2 image-generation model, was rolled back less than a day after its public debut following demonstrations showing how it could generate fabricated images of wars, natural disasters and sensitive military installations.

Why Google Pulled the Feature

The AI tool allowed users to generate or modify realistic satellite-style images of real-world locations using text prompts.

While Google said the technology was intended to support education, urban planning and geospatial analysis, users quickly demonstrated its potential for misuse by creating fictional scenes of bomb damage, refugee camps, flooding and other events that never occurred. These examples sparked concerns that the tool could undermine trust in satellite imagery, which is widely used by journalists, researchers and humanitarian organisations to verify real-world events.

In response, Google confirmed that it had paused the feature while it develops stronger safeguards against abuse.

Google Responds

Google said the AI-generated images were never incorporated into the standard Google Earth experience and were marked using SynthID, the company's digital watermarking technology for AI-generated content.

However, the company acknowledged that additional protections are needed after users shared screenshots of manipulated images outside the platform.

"We're temporarily rolling back this experimental feature while we implement stronger safeguards," Google said in a statement.

Experts Warn of Misinformation Threat

Cybersecurity researchers and OSINT analysts argued that integrating generative AI into a trusted mapping platform creates new risks because fabricated satellite images could be mistaken for authentic evidence.

Several demonstrations showed how prompts could generate realistic images depicting fictional military bases, infrastructure damage or humanitarian crises, potentially complicating efforts to verify events during conflicts or natural disasters.

Although Google maintained that AI-generated images contain digital watermarks, critics argued that watermarks alone may not be sufficient once images are widely shared across social media platforms.

Balancing Innovation and Trust

The incident highlights the growing challenge facing technology companies as increasingly realistic generative AI tools become available to the public.

Google said it remains committed to developing AI features for Google Earth but will only relaunch the image-generation capability after introducing stronger safety measures designed to reduce the risk of misinformation and policy violations.

Broader AI Safety Debate

The rapid withdrawal of the feature has intensified discussion about how companies should deploy powerful generative AI systems, particularly those capable of producing realistic content linked to trusted information platforms.

Industry experts say the episode underscores the need for stronger technical safeguards, transparent labelling and careful testing before advanced AI tools are released to the public, especially where they could influence public understanding of real-world events.