Sunday, July 5, 2015

Do We Need A Google "Truth Algorithm" For Web Searches?

Back in March, Google announced a new search algorithm that would display websites and news stories based on their factual content at the top of the list and the factually-deficient results at the bottom.  This is radically different from current practice where Google puts the top read results at the top and less read to the rear. But is this the way it should be, given people may be reading too much bollocks and not enough factual, scientific material?

Should creation science websites or climate denier sites be at the top of searches because  most people read them, or at the bottom because they are not factual? If the former is acceptable, then really all we have is a popularity contest. Is that valid for those seeking genuine information as opposed to PR and nonsense?

Interestingly, when Google first announced this new search program, there was instant blowback and outrage from FOX News (which consistently ranks low in accuracy at Politifact.org and Factcheck.org.) Why would FOX be howling its head off? Maybe, just maybe because their honchos know most of what they put out there is gibberish and distortions. From O'Reilly to Hannity and Fox n' Friends, none of it can be trusted.

FOX wasn't the only bunch screaming. Others included: climate change deniers, Vaxers (those against childhood vaccinations), and birthers.  Again, why are they howling like stuck pigs if they are offering valid information? There'd be no reason for such outbursts.

Sadly, Google has no plans to start using the program but it's interesting to consider the uproar if they did given all the liars, PR-mongers, disinformationists and others would have to come out of the woodwork. They'd not sit still or remain quiet given it might jeopardize their interests.

But think of the benefits!

For example, employers could fact check job candidates' resumes and LinkedIn pages, identifying bogus credentials including inflated education credentials that might eliminate many job seekers on the basis of fudged resumes.

Even more importantly, kids and others interested in advancing their educations on the subject of climate change, evolution and American history wouldn't be driven first to nonsense sites such as those of Roger Pielke Jr., Ken Ham and the history revisionist Larry Schweikart, see e.g.

http://brane-space.blogspot.com/2014/08/48-liberal-lies-how-about-50-tea-party.html

Imagine how much more efficient educational web searches would become if all the detritus could be instantly ranked at the bottom instead of the most relevant and factual material - forcing truth seekers to plow through page after page, even after refining searches.

But it seems this is what we will have to settle for until Google implements its truth algorithm. That means facts and truth will not come easily but only after diligent work. Of course, in this quest reading can also do wonders, because it upends the popularity ranking systems prevalent for existing web search engines.

Point of fact, babble such as put forward by the likes of  Larry Schweikart, Ken Ham or Roger Pielke Jr. wouldn't stand a chance at being seen as credible if people read more diligently from the pool of books shown to be factual or scientific. Books such as Darwin's Origin of Species, and James Loewen's tour de force ('Lies My Teacher Told Me') exposing the bogus American history taught our kids- but more importantly also disclosing the actual history omitted!

Moral of the story: read first, or at least read to check any web searches you pull up - don't just accept the first ones as the factual ones.

2 comments:

  1. My instinct on reading this was that a general truth algorithm would be a very difficult problem, almost as hard as (and possibly even equivalent to!) a human-level AI. This site seems to agree that, whatever it is, it's not here yet:

    http://www.searchenginejournal.com/googles-truth-algo-5-facts-know/129103/

    ReplyDelete
  2. I tend to agree in that at multiple levels ascertaining truth will be subjective, based on "official" versions already circulated and on which are conferred the "benediction" of verity. For example, it is almost a given that any such algorithm now would regard the Warren Commission version of the JFK assassination as "the truth" - historically - despite the fact deep research will reveal it as a poorly assembled tissue of lies, omissions and major distortions - more a fraud than history. But because of its mistaken "gravitas" any overly restrictive truth algorithm will tilt toward it. The same applies to numerous science areas, i.e. are black holes real or cosmic strings?

    ReplyDelete