It's getting harder to distinguish real from fake again: Google allows removal of watermarks on AI images

Distinguishing between real and AI-generated images is becoming more challenging as Google introduces an option for users to remove the visible watermarks that identify AI-created content.

The 'sparkle,' a four-pointed star icon previously used by Google to indicate AI-generated images, is no longer mandatory. This change follows user feedback, where the visible watermark was cited as a significant annoyance. Google's move aligns with competitors like ChatGPT, which also allows for the creation of images without visible watermarks.

This development raises concerns, especially after an AI-generated image of Jeffrey Epstein circulated widely on social media, falsely claiming he had faked his death. The original image contained the Gemini watermark, but versions that went viral had it deliberately removed, highlighting how the absence of visible markers can facilitate the spread of misinformation.

While Google is removing the visible watermark, the company is still obligated to embed invisible watermarks in AI-generated content. European regulations, for instance, require watermarking on AI images, particularly when there is a risk of deception. This invisible marking can still be detected by tools like Gemini or Google Search, offering a way to verify content even when the visible signs are absent.