tech
It’s Make or Break Time for AI Labeling Systems
Posts from this topic will be added to your daily email digest and your homepage feed.
TL;DR
- Google is integrating its invisible watermarking system, SynthID, into Chrome and Search, expanding its reach significantly.
- OpenAI will now embed SynthID watermarks into images generated by ChatGPT, Codex, and its API, alongside existing C2PA metadata.
- The expansion of SynthID and C2PA aims to combat the spread of deceptive AI-generated content online.
- C2PA metadata can be easily removed, raising concerns about its reliability as a sole solution for content provenance.
- SynthID is considered more robust as its markers cannot be easily stripped out, making it a potential tool for debunking deepfakes.
- Meta plans to use C2PA metadata to tag images captured by cameras on Instagram, aiming to help users differentiate real photos from AI fakery.
- The success of these labeling systems hinges on widespread adoption by AI model developers and online platforms, as well as user trust and effective implementation.