Takedown

Our takedown process will occur in two phases: recognition and escalation.

Based on our watermark, we run a machine learning classification algorithm to search for our unique watermark noise by scraping existing internet content. If a match is found, we follow established processes to escalate DMCA requests for out-of-scope content on the creator’s behalf.

As our content verification network grows, we begin to form a large-scale in-scope/out-of-scope definition for users, domains, and content; in this, we are able to more effectively identify exceptions to user-defined scope rules and unauthorized content distribution. By partnering with prominent CDNs serving content to the majority of the internet (Cloudflare, Stackpath, Rackspace) our scoped approval network will become an embedded step to serving symetrically-controllable content en masse.

We describe an additional takedown mechanism in our Tokenomics section where members of the community run public models for detection (not unlike the Folding@home distributed computing project).