On Friday, July 31, 2026, Google withdrew a recently introduced feature on its Google Earth platform that enabled users to generate hyper-realistic satellite images using Nano Banana, Google's own AI image generator. The feature had been live for just over 24 hours and allowed anyone to create images depicting fabricated events, such as destruction at Singapore’s Changi Airport or a fire at Gardens by the Bay, with simple text prompts.

Initially, Google presented the feature as a tool for imagination and planning, such as visualizing new neighborhoods or real estate projects. However, it quickly faced criticism from experts in journalism, artificial intelligence, and open-source intelligence who warned that the ease of creating such images could accelerate the spread of disinformation.

Prominent open-source investigator Henk van Ess was among the first to highlight the risks, stating in a blog post that the feature could cause more harm than good. Tom Jarvis, an AI specialist from the Centre for Information Resilience, commented on LinkedIn that "Google has made a feature that turns the tables for the worse," noting that those accountable through satellite imagery could now disrupt information spaces more easily.

Google Earth acknowledged in a statement that while geospatial professionals used the feature for various legitimate purposes, they had observed users sharing generated images that violated their policies. The company emphasized the unique trust people place in Google Earth for reliable world views.

Benjamin Ang, head of the Centre of Excellence for National Security at Singapore's Rajaratnam School of International Studies, told CNA that satellite images have historically been seen as more credible than ground-level photos due to their objective angle and difficulty to tamper with. However, he warned that once such images circulate in family or friends chat groups on platforms like WhatsApp, Telegram, or WeChat, few people verify their authenticity before sharing.

Experts also noted that creating convincing fake satellite images previously required significant resources and expertise, but the new AI tool lowered these barriers, enabling almost anyone to produce misleading images that could cause panic or outrage.

The rapid removal of the feature underscores ongoing challenges in balancing technological innovation with the risks of misinformation in the digital age.

Sources