Following the initial work and discussions around Blockcerts' Certificate revocation there was a feeling that a similar system (whith a lot of similar parts) could be used for something around Fake News mitigation.
On a recent meeting with prof. Nuno Santos, we came to the conclusion that a way to harmonize these two efforts - Blockcerts' Certificate Revocation and Fake News Mitigation - could be built under the umbrella of a Verifiable Claims manager. Essentially both those implementations require Verifiable Claims (which are nothing more that non-repudiable statements, whith assurances of integrity) with a revocation method to work properly.
The Fake News effort revolves around Clickbait detection, that is verifying that a news article's title is in accordance with the news article's body. The core component of that system is the act of a Verifier (someone who verifies news articles) creating a claim (Verifiable Claim for instance) that states the validity (or lack of) of a news article's title.
Blockcerts' certificates are also evolving towards a model that uses verifiable claims (while maintaining compatibility with Open Badged). The positive implication of this is that we can develop a way to revoke verifiable claims we automatically have a way to revoke Blockcerts.
These two updates made clear that the link between the, let's call it Clickbait detection, and Certificate Revocation efforts was Verifiable Claims, and that is how Hypercerts went from being a plug-in for Blockcerts to a fully fledged Verifiable Claims system.
The system's core functions are areissue-veredict, performed by a Verifier and check-veredict, performed by a Reader.
The issue-veredict function is the action taken by a verifier (a user that analyzes a news article's content and asserts wether the news title is accurate) which results in the issuance of a verifiable claim, whose contents inform reader about the correctness of a news title.
The function is called by a verifier, which passes to it his identity, V, the news article's title, title, the news article's body, body and the URL link to the article hosted by the news outlet, URL.
The function outputs an ack to let the verifier know that everything went accordingly. Internally the system is generating and storing a verifiable claim, as well as the necessary proofs that go along with it.
The check-veredict function is called by a reader who wants to know wether a certain news article's title is accurate. The function is fed the URL of the news article and returns the verifiable_claims generated by verifiers, for that article.
The building block of the Clickbait Detection system is Verifiable Claims. Thus, the system assumes three Verifiable Claims' (VC) primitives:
- Create VCc
- Verify VC
- Revoke VC