DRAFT Credible Web: Solution Landscape
Approaches
Potential approaches to increasing web credibility via data sharing
Inspection
Self-Report (1st Party)
- See The Trust Project
- Transparency, self-certification
- Data feed vs UI
Annotation (3rd Party)
- See Credibility Coalition
- Consider difficulty of faking (“gaming”) signals
- Consider explicit/conscious vs implicit/incidental signals
- paid vs volunteer, friend vs stranger
- Use of Web Annotations
Corroboration (Claim-Based Systems, Fact-Checking)
- See IFCN, ClaimReview
- Detecting and sharing checkable claims in content
- Distributing demand for checking
- ClaimReview spec details
- Presentation conventions (CrossCheck, Annotation Layer)
- Matching/sharing repeated/similar appearances
- Tagging evidence
- Tagging claim relations (refutation, broader/narrower)
- Change over time, change in context
Reputation (Trust Networks)
- Allow user to pick their roots of trust, making a “personal walled garden”
- Help users navigate and assess trustworthiness of sources
- Trust = Accuracy (truth) + Benevolence + Competence
- Bootstrap from: followers? co-authoring?
- Danger in asymmetry, as negative info seen as more salient
- Danger in coerced statements (eg need for secret ballot)
Out Of Scope
- Specifying quality metrics or certification standards for journalistic practice or other content providers
- Assessment of credibility content or sources, beyond examples and test data used for development and education
- Developing algorithms or constructing machine-learning training sets to enable machines to assess credibility of content or sources
- Any technology which disempowers end-users or is likely to lead to increased censorship
- New data formats, APIs, or cryptosystems. (Instead, we will support reuse of existing standards like JSON-LD.)
- Assessing “hate speech” and other “conversational health” issues outside of credibility
- JSON-LD
- Web Annotations
- SPARQL
- PROV Model