Google has removed its new Google Earth AI satellite image generation feature less than a day after launching it, following criticism from open-source intelligence (OSINT) researchers who warned that the technology could be used to create realistic fake satellite imagery and spread disinformation.

The company rolled back the “Create Image” feature within 24 hours of its release after experts raised concerns about its potential impact on public trust in geospatial data. The tool used Google’s Nano Banana 2 image generation model, allowing users to zoom into any location on Google Earth and create synthetic satellite, aerial or 3D images simply by entering a text prompt.

Google said the feature was introduced to support creative and professional uses but decided to suspend it temporarily after users shared AI-generated images that appeared to violate the platform’s policies.

Trump’s Gaza Peace Board Assures Netanyahu Israel Won’t Withdraw Before Hamas Disarms

According to the company, all generated images were clearly watermarked as AI-generated and never appeared in the main Google Earth experience viewed by the public. Google stressed that the rollback is temporary while engineers strengthen safeguards against misuse.

“We know that people uniquely trust Google Earth for a reliable view of the world,” the company said in a statement. Google added that although geospatial professionals had found valuable uses for the tool, some generated images shared online raised concerns about policy violations, prompting the company to introduce stronger protections before relaunching the feature.

The decision followed criticism from members of the OSINT community, who argued that highly realistic AI-generated satellite images could be exploited to spread false information during conflicts, disasters and geopolitical crises.

Open-source researcher Henk van Ess demonstrated the issue by testing prompts involving refugees at the Mexican border, a nuclear facility in Iran, an aircraft crash in Amsterdam and a hospital with a bomb crater in Gaza. He said the system generated images for all of the prompts without rejecting them.

Brady Africk, a research analyst at the American Enterprise Institute, warned that convincing AI-generated satellite images could undermine confidence in authentic imagery. He said such content could make it significantly harder for journalists, investigators and OSINT researchers to verify real-world events and identify manipulated visuals.

The rapid rollback highlights the growing challenge facing technology companies as they expand generative AI capabilities while trying to prevent misuse. It also underscores increasing concerns over the role of AI-generated content in misinformation campaigns, particularly when realistic images can influence public perception during breaking news and international conflicts.