Google Suspends 'Nano Banana' AI Tool After Deepfake Satellite Alarms

Alphabet has officially suspended its newly launched AI image generator, known as Nano Banana, from the Google Earth platform just 24 hours after its public release. The decision follows widespread reports of users generating hyper-realistic deepfake satellite imagery and synthetic landscapes using simple text prompts. Investigations revealed that the tool could be used to fabricate geographical features, environmental scenes, and urban events, potentially compromising the integrity of satellite data relied upon by researchers and journalists.

The incident highlights severe risks regarding the integrity of global mapping data and digital evidence. Open-source intelligence investigators, human rights groups, and news organizations use Google Earth to verify human rights issues, environmental changes, and military infrastructure. By allowing the superimposition of AI-generated objects onto real-world coordinates, the tool inadvertently facilitated the creation of deceptive visual information, including fabricated disasters and altered border regions circulating on social media.

Google stated that it chose to pause the feature to prevent further misuse while refining safety filters, watermarking technologies, and verification mechanisms. This development carries major implications for national security, particularly in sensitive regions where borders are high-stakes like India, where geographical data is critical for security and tracking development. Moving forward, the industry faces mounting pressure to establish mandatory watermarking and advanced authentication standards to prevent the weaponization of spatial data.