@stefanOn that page, they are indicating that their biggest financial challenge is their "Content Classification Service (CCS)", a hash-and-match "solution" that is mostly promoted for #CSAM identification. They're also mentioning reporting the respective content to authorities, if found:
»We integrate with your instance using webhooks, processing content you send us to hash the media and match it with known CSAM. Your media never leaves IFTAS. If we find pertinent matches we’ll provide human review, notify you via email and issue a takedown request. We perform any required reporting and law enforcement record-keeping so you don’t have to.«
In the course of the #ChatControl debate in the #EU, it has been widely discussed how such hash-and-match "solutions" can be extended to detect any kind of content, simply because those in control of the hash databases control what gets detected and reported..
IFTAS themselves say: »Future classification services will include hash and match options for non-consensual intimate images, terroristic and violent extremism content, spam, and more.«
In short, the Four Horsemen of the Infocalypse are riding again.
They also say »We cannot open the underlying hash matching databases«, which means: no local hash databases, hence remote hashing is required.
I'd think twice if I really want to pass on all content of an instance to an external entity, for hashing and and potential reporting. We are talking about "China and dissident content" situations here, but in the context of the USA of 2025 and beyond – I'd certainly not want to donate to help implementing this.
@iftas
