Social media platform Snapchat recently unveiled plans to introduce watermarks on images produced through artificial intelligence tools, in response to the surge in AI-generated visuals. The addition of these watermarks will help users identify images created using Snap’s generative AI tools.
These watermarks will feature a ghost logo alongside a recognizable sparkle icon, indicating that the image was AI-generated on Snapchat. The watermark will be visible when the image is exported or saved to the camera roll.
In line with other companies, Snapchat has incorporated AI-driven features for its premium subscribers to edit or enhance photos, some of which may not be entirely AI-generated. To provide users with transparency, these edited images will also carry a watermark to signify the use of AI technology.
Snapchat utilizes contextual icons and labels within its app interface to clarify to users when they are interacting with AI-powered features. For instance, when sharing an AI-generated Dreams image, recipients will see a context card with additional information. Features like the extend tool, which utilizes AI to zoom out snaps, are distinguished as AI features with a sparkle icon for users creating the content.
Unlike certain other platforms, Snapchat’s watermarking approach is distinctive. Google and Dall-E, for instance, introduced watermarks for AI-generated images that do not alter the visual appearance but are embedded in the metadata.
Similarly, Meta, the parent company of Instagram and Facebook, disclosed plans in February to label AI-generated images across its platforms. Moreover, Meta is not solely singling out images produced using its own AI model but is also identifying AI-generated images comprehensively.
Snapchat and Meta’s initiatives to label AI-generated images arose partly from concerns related to political transparency, especially with the approach of the next U.S. presidential election. Both platforms aim to ensure transparency in political advertising by subjecting all political ads to a thorough human review process.
This scrutiny includes checking for any misuse of content, such as the deceptive manipulation of images or content using AI technology, as emphasized in Snap’s statement.