tech
Google Earth's AI deepfake tool only lasted one day
Posts from this topic will be added to your daily email digest and your homepage feed.
TL;DR
- Google Earth launched a new AI feature allowing users to edit satellite images with text prompts.
- Users could create AI deepfakes, with examples including refugees near the Mexican border and a bomb crater in Gaza.
- Despite watermarks and Google's policy against harmful topics, users found the tool easy to exploit.
- One user was able to fool an AI detection tool with a generated video.
- Google has rolled back the feature to implement stronger safeguards, stating it was not visible to others and was watermarked.