IMO, the typical approach of using fact-checking services to rate the accuracy of sources is inevitably flawed: if a source (or a fact checker) builds a reputation for reliability, it will eventually be suppressed or subverted into exploiting its reputation for other purposes.
A better option might be to treat all sources as potentially informative, but not at face value: rather, build a predictive model of each source, and treat as significant only those stories that deviate from prediction (i.e., stories that seem atypical for that source). Those are the stories most likely to convey information the source didn’t generate itself.
That’s certainly a good point, but I’m less concerned with how to verify information than with how to counteract the constant flow of misinformation — especially on other platforms where misinformation is deliberately pushed, which is causing major problems in my home country alone.
IMO, the typical approach of using fact-checking services to rate the accuracy of sources is inevitably flawed: if a source (or a fact checker) builds a reputation for reliability, it will eventually be suppressed or subverted into exploiting its reputation for other purposes.
A better option might be to treat all sources as potentially informative, but not at face value: rather, build a predictive model of each source, and treat as significant only those stories that deviate from prediction (i.e., stories that seem atypical for that source). Those are the stories most likely to convey information the source didn’t generate itself.
That’s certainly a good point, but I’m less concerned with how to verify information than with how to counteract the constant flow of misinformation — especially on other platforms where misinformation is deliberately pushed, which is causing major problems in my home country alone.
How are you going to counter misinformation if you can’t determine what is and isn’t misinformation?