A few weeks ago we discussed the idea of credibility filtering as the future of news filtering. In the age of constant information in which we live, filtering, we said, is surely the future of news. How to best sort out the signal from the noise is still to be determined.
The two most popular approaches right now — crowd-based human filtering, and algorithmic filtering — are not without their problems. As we discussed in August, filtering based on the credibility of the source is an idea that is gaining some steam. We reviewed two startups that filter news using a combination of all three approaches — crowd-based, algorithmic, and credibility. The problem with their approach, we said, was that the definition of credibility varies from person to person.
Today, semantic search engine Hakia announced that it is jumping on the credibility bandwagon by offering a way for users to restrict results to so-called “credible” sites.
Credible sites will be reviewed by “librarians and information professionals” and hand-picked to meet specific credibility requirements. Those requirements present a major problem, though. Hakia defines a credible site as one whose content is peer-reviewed, that has “no commercial intent or bias,” that is current, and whose publisher can be verified.
The no commercial intent requirement is a sticking point for me. That instantly rules out a lot of credible, well-researched media sources simply because their content is created to drive revenue to the publisher. It also means that many non-profit or governmental sources might be included even though their content includes a clear bias (just, perhaps, not a commercial one).
Filtering by credibility is a good idea — being able to weed out the credible results from the rubbish would be great, but Hakia’s approach is flawed, in my opinion. It will, by definition, pass over a large number of perfectly credible results, and potentially — and dangerously — include results that are actually less credible.
Credibility filtering is perhaps where social search might come in handy. If a search engine, like the two news sites we reviewed in August, relies on the crowd to rate the credibility of the source, and you trust your friend’s definition of credibility, the search engine could start returning results based on a customized library of sites that you and your friends have deemed the most credible. For those without a large group of trusted peers, it could suggest “people like you” after you’d begun to tag and rate sites based on credibility.
None of these approaches is perfect, but figuring out how to reliably and accurately filter all the content we’re creating will be the next big thing on the web — perhaps the next “Google” opportunity.