Social media company Snapchat announced its plan to add watermarks to AI-generated images on its platform. The company intends to include a logo of a small ghost with a sparkle icon to denote an AI-generated image. This watermark will be visible when the user exports the image or saves it to the camera roll.
Enforcement of Watermark
Snap stated on its support page that removing the ghost with sparkles watermark violates its terms of service. However, it remains unclear how Snap plans to detect the watermark removal.
Industry Trends in AI Watermarking
Snap’s decision aligns with the broader trend among tech giants to label or identify images created with AI-powered tools. In addition, companies such as Microsoft, Meta, and Google have also taken steps in this direction.
Challenges and Criticisms
The implementation of watermarks for AI-generated content is a step towards combating misinformation. However, there are challenges and criticisms associated with this approach. Some experts argue that AI watermarking may not be effective in curbing disinformation. Additionally, it could hinder attempts to automatically label AI-generated content at scale.
Snapchat AI-Powered Features and Safety Practices
Currently, Snap allows users to create or edit AI-generated images with Snap AI for paid users. In addition, it has a selfie-focused feature called Dreams.
In a blog post, the company outlines its safety and transparency practices around AI. In addition, Snap explained that it shows AI-powered features such as Lenses with visual markers like a sparkling logo. Additionally, they added context cards with AI-generated images from tools like Dream Selfies to better inform the user.
Snapchat Efforts to Ensure Equitable Access and Minimize Bias
In February, Snap partnered with HackerOne to stress its AI image-generation tools by adapting a bug bounty program. The company also created a review process to remove problematic issues when AI-powered lenses are in development. In addition, the firm emphasized its commitment to implementing additional testing to minimize potentially biased AI results.
Addressing Past Issues:
Snapchat faced challenges with its AI-powered features in the past. Additionally, it includes the “My AI” chatbot returning inappropriate responses to users. In response, the company rolled out controls in the Family Center for parents and guardians to monitor and restrict their teen’s interactions with AI.
Also Read: https://thecitizenscoop.com/x-plan-to-introduce-fees-for-new-users/
