Google Earth's AI Image Generation Function Pulled in Less Than a Day

Google pulled its AI image-editing feature from Google Earth within 24 hours of its launch, after users created deepfakes depicting disasters.

Google launched the AI image-editing feature on Thursday, allowing users to access Google Earth and transform locations with text prompts using Gemini's Nano Banana 2 image model. The feature was available for less than 24 hours before being pulled due to users creating deepfakes depicting disasters. According to reports, the feature was accessed by thousands of users during its brief availability. The decision to pull the feature was made by Google's product development team, led by Vice President Jen Fitzpatrick.

The removal of the AI image-editing feature from Google Earth directly affects users who rely on the platform for geographic information and research. For instance, researchers using Google Earth for environmental monitoring or disaster response planning may face limitations due to the lack of advanced image editing capabilities. This limitation can impact the accuracy and effectiveness of their work, potentially affecting the quality of research and decision-making. As a result, these users may need to seek alternative platforms or tools.

The incident is part of a larger trend of tech companies struggling to balance innovation with responsible AI development. In recent years, companies like Facebook and Twitter have faced similar challenges with AI-powered features, highlighting the need for more robust content moderation and safety protocols. Insiders familiar with Google's AI development process note that the company has been working to improve its content moderation capabilities, but the rapid launch and removal of the feature suggest that more work is needed. This incident may influence the development of future AI-powered features.

Google is expected to announce an updated version of the AI image-editing feature in the coming weeks, with enhanced safety protocols and content moderation measures. The company's upcoming product development conference, scheduled for March 15, may provide more information on the feature's future. Notably, Google's decision to pull the feature has sparked a debate about the role of AI in content creation, with some experts arguing that the technology is not yet ready for widespread use. Interestingly, the Nano Banana 2 image model used in the feature is still available for use by developers, potentially allowing other companies to integrate similar capabilities into their own products.