OpenAI launches cryptographic marking system to verify AI images
Technology
OpenAI just launched a smarter way to tell if an image was made by AI.
It's now using SynthID, a fancy word for hiding cryptographic information inside pictures so you can't see it, but it's still there even if someone edits the image.
This makes it much easier to check if OpenAI-generated images are truly AI-generated.
OpenAI adds invisible DeepMind SynthID watermarks
OpenAI is also incorporating Google DeepMind's SynthID to add invisible watermarks to images from ChatGPT and its API.
These marks survive common tweaks like cropping or color changes, making verification simple across platforms.
Plus, OpenAI is sticking to C2PA standards for secure metadata, so spotting real vs. fake content just got a lot more reliable.