SnapChat intends to incorporate watermarks into images generated by its AI-powered tools, according to an announcement made on Tuesday. These watermarks will feature a translucent version of the Snap logo accompanied by a sparkle emoji and will be applied to any AI-generated image exported from the app or saved to the device’s camera roll.
The company has stated that removing these watermarks from images would breach its terms of service, although it has not specified how it plans to detect such removals. SnapChat joins other major tech companies like Microsoft, Meta, and Google, which have implemented measures to label or identify images created with AI-powered tools.
Snapchat currently offers paying subscribers the ability to create or modify AI-generated images using Snap AI. Its “Dreams” feature, which focuses on selfies, empowers users to enhance their photos with AI.
In a blog post detailing its safety and transparency practices regarding AI, Snap highlighted its use of visual markers resembling a sparkle emoji to denote AI-powered features such as Lenses. Additionally, the company has introduced context cards to provide users with more information about AI-generated images produced with tools like Dream.
In February, Snap initiated a bug bounty program in collaboration with HackerOne to rigorously test its AI image-generation tools. The company emphasized its commitment to ensuring equitable access and expectations for all users, particularly regarding AI-powered experiences, by conducting additional testing to mitigate potentially biased AI outcomes.
Snapchat’s endeavors to enhance AI safety and moderation follow the launch of its “My AI” chatbot in March 2023, which sparked controversy when some users prompted the bot to engage in discussions about sensitive topics like sex and alcohol. Subsequently, the company introduced controls in the Family Center, allowing parents and guardians to oversee and restrict their children’s interactions with AI.
