Blog Posts

Knowledge Based Trust?

Thoughts on Google’s project to rank websites downgrading them on its search engine based on the number of ‘false facts’ they contain. Proviso. The opinions contained are entirely my own, not necessarily those of Razormind, and are of course subject to debate.

Are there a sufficientnumber high ranking inaccurate websites to necessitate such a move by Google? Google already downgrades sites that link to each other, are not accurately titled, are spamming and so on. Currently Google ratings are based on several factors, including the number of links to a website, assuming that popularity denotes trustworthiness.Is this new false facts criteria just an extra layer or a replacement for current criteria?

The paper the Google project is based on is linked here. http://arxiv.org/abs/1502.03519v1. What would the outcome of the method described within be likely to be?

An interesting blog post and comments on this topic is https://plus.google.com/+RandFishkin/posts/TAKxDuwvbXD

IMHO. Pro- There is plenty of misinformation out there that never gets debunked no matter how much evidence to the contrary there is. This is unlikely to be solved by more evidence as the beliefs appear to be emotionally held rather than rational. http://www.nobeliefs.com/beliefs.htm

Against- How does Google decide what is a fact? Currently this appears to be if it is agreed by 90% of the net. Would this decision, which will be overseen by humans, apply only to verifiable facts such as addresses or also to opinions and theories? If 90% of the net believing a fact makes it so, as far as Google is concerned, what would the affect be on currently unpopular, controversial opinions? For example Electric Universe theorists. http://www.electricuniverse.info/Electric_Universe_theory Would the charge of ‘pseudoscience’ be enough to have a site disappear from Google rankings?

Where and how would Google get its facts? https://www.cs.cmu.edu/~nlao/publication/2014.kdd.pdf If Google takes its facts from its Knowledge Vault are facts from government sources and security services, such as the CIA Fact Book, considered trustworthy facts and therefore ranked highest? Could this even be a way for Google to decide both the nature of facts and what the facts are and would that amount to censorship on a tremendously Orwellian scale?

Is having false facts a part of the democratic nature of the web? Should people be able to analyse and decide for themselves what is true rather than have truth handed to them pre-packaged? Is it a good idea to shut out those who have minority/different/annoying/offensive opinions, to us,rather than have their opinions freely available,even if we consider them to be wingnuts? What if people consider us to be wingnuts? What about the entertainment value of wingnuts? I want to be able to read about Flat Earthers!

Personally, I’d rather have an optional filter on celebrity gossip than on false facts. If Google wants less websites with dubious facts highly ranked is there another way of dealing with it? Are sites with dubious facts really such a menace? More than advertising, spam, mainstream media etc?

What do you really want Google? Looks like being an Oracle is a higher aspiration than being a search engine and Facebook, Amazon etc are racing to build their own knowledge vaults. http://www.newscientist.com/article/mg22329832.700-googles-factchecking-bots-build-vast-knowledge-bank.html#.VUJ1oSFViko These will pull in data about all web users as well as other kinds of data. Kurzweil predicts everyone on earth will be able to jack straight into the cloud, thereby presumably not just able to access all info but having all their data accessed too. Neuromancer? http://www.techrepublic.com/…/kurzweil-your-brain-will-con…/ Comments , corrections and tech perspectives welcome.