Google Wants To Rank Websites Based On Facts Not Links

wabrandsma writes about Google's new system for ranking the truthfulness of a webpage. "Google's search engine currently uses the number of incoming links to a web page as a proxy for quality, determining where it appears in search results. So pages that many other sites link to are ranked higher. This system has brought us the search engine as we know it today, but the downside is that websites full of misinformation can rise up the rankings, if enough people link to them. Google research team is adapting that model to measure the trustworthiness of a page, rather than its reputation across the web. Instead of counting incoming links, the system – which is not yet live – counts the number of incorrect facts within a page. 'A source that has few false facts is considered to be trustworthy,' says the team. The score they compute for each page is its Knowledge-Based Trust score. The software works by tapping into the Knowledge Vault, the vast store of facts that Google has pulled off the internet. Facts the web unanimously agrees on are considered a reasonable proxy for truth. Web pages that contain contradictory information are bumped down the rankings."

Comments

YES

[Anonymous Coward]

It's about time, but I really hope their 'factual accuracy' engine gets open sourced so we can be clear on exactly how they determine what are 'facts'

Re:YES

[ShanghaiBill]

It's about time

Yes, automated groupthink is a wonderful idea. Doubleplusgood!

Re:YES

[lgw]

DuckDuckGo has ranted for a while about Google's "search bubble", where it shows each user pages likely to confirm their biases. This twists that old habit, so now everyone will be presented with pages likely to confirm Google's biases. Disagree with the Google groupthink? Your page is filled with lies, and Google will do its best to hide it.

Your own good judgment is the only worthwhile filter, and you don't get or maintain that by seeing only pages that all say the same thing. You don't really understand any subject until you can argue both sides in detail, and see why those with the out-of-favor view believe what they do!

I don't expect it will take very many years before Google picks sides on politically contentious issues in their ranking. Is abortion murder? Don't worry, Google will decide for you! Is recycling actually helping the environment? Don't worry, Google will decide for you! Are the current Net Neutrality changes actually good for the consumer, or only for internet giants? Don't worry, Google knows where the facts are!

Whatever you do, don't spend the time to study issues in depth for yourself, no, don't rise above your station. Repeat what you've been told, and everyone will say how smart you are - well, at least everyone who shows up in a Google search.

Bad move

[Whiteox]

That WILL be a bad move. There are a lot of facts out there that academics still debate over. Pretty much anti-free speech afaic.

Re:Bad move

[hey!]

It is seldom the veracity of facts that the debate is over; it is their significance. But that happens to be where this falls idea falls short, because misinterpretation of facts is where the most potent misinformation comes from.

Case in point, "vaccine injury" -- which is a real thing, albeit very rare. Anti-vaccine activists point to the growing volume of awards made by the US "Vaccine Court" (more accurately called "The Office of Special Masters of the U.S. Court of Federal Claims") as proof that vaccine injuries are on the rise.

It is a verifiable fact that the volume of awards has grown since the early years of the program. That is absolutely and unquestionably true. However, that this is proof vaccine injuries is gross misinterpretation, because the "Vaccine Court" program is no fault. You don't actually have to show the defendant *caused* an "injury", you only have to (a) show the child got sick after being vaccinated and (b) find a doctor to sign off on a medical theory by which the child's illness *might* have been caused by the vaccination.

Since you don't have to actually prove injury in "Vaccine Court", the rise in cases and awards doesn't know vaccine injuries are on the rise. All that is necessary is that more people think that their child's illness was caused by vaccinations, and the low burden of proof will automatically ensure more awards.

And so there you have it. A perfectly factual claim can be cited in a way that leads people to preposterous conclusions.

can't wait to see it work on fox news web site

[youn]

It is an interesting algorithm, but who is to say the fact is correct? if many sites say the same information, the fact is more correct? if that is the case, then how much better is that from links. when there is more than one version of the truth (conflicts, spin vs fact)... plus not all information is facts... philosophical questions may have more than one answer etc... so I am definitely curious to see how this works out

Re:can't wait to see it work on fox news web site

[sumdumass]

Not in science. A fact is an observation or evidence that has been repeatedly observed to be true. It doesn't mean always true or only true.

The problem is when existing theories compete. OR more precisely points within large theories compete. Take relativity for instance, gravitational waves help explain the big bang but not all observations support the big bang model. But gravitational waves are considered fact for the purpose of the theory even though it has never been directly observed because it can be explained in mathematical computations that explain observations.

So what happens when we actually detect them for real and they operate slightly different than we think? Does this new observation or fact get pushed to the front of the line or is it buried because the fact engine hasn't updated yet or the wikipedia article it is referencing is in a mod battle. How about if something else is found to explain the theory concerning gravitational waves but lives in the same limbo as gravitational waves in which it hasn't been directly observed but can explain observations with math also.

It reminds me in the 80's when (and I forget who) some doctor was claiming most stomach ulcers were the result of bacteria. Turns out that is a fact but he was originally ridiculed because the fact at the time was that no one believed that bacteria could survive in the stomach's acidic environment longer than it takes to pass through it. Now the fact is that it's cheaper to just giving a couple antibiotics and seeing if the ulcer disappears than to test if the ulcer is bacteria related or other. But it was indisputable at one time, then someone disputed it and now it is indisputable again. Facts change.

Re:can't wait to see it work on fox news web site

[Harlequin80]

The stomach ulcer thing was two Australia scientists, Barry Marshall and Robin Warren. They claimed, in 1982, that stomach ulcers were caused by H. Pylori but because that bacteria was common in the intestinal tract already the idea was shot down as incorrect. They spent 2 years trying to infect piglets with it then gave up and Barry drank a full petrie dish of the cultured bacteria. He suffered symptoms within 3 days, and had significant ulcers and gastritis after a week.

As a result of this test and the observed symptoms the medical community accepted their findings. They were awarded a nobel prize, in '96 I think.

Re:can't wait to see it work on fox news web site

[Strangely Familiar]

The Google policy is about facts, not things. Facts are representations about things. Actual things, for example, cannot be true or false, they just are. Facts which represent properties of those things can be true or false. The Empire State Building is a particular thing. It is not true or false. "The Empire State Building is painted red" is a false fact about that particular thing. So, before anybody said, "H. pylori cause ulcers" it did not exist as a fact, only a thing. Remember, a fact is a representation, not to be confused with the thing it represents. False facts about the cause of ulcers existed. Few knew them to be false, except for the two Australians. The point that the GP was making was that Google doesn't deal in things, it deals in representations (web sites) about representations (other web sites). It doesn't do experiments on people to see if bacteria acutally cause ulcers; it relies on other peoples' representations about that. So when you say, "The concensus changed, not the fact that bacteria caused the ulcers." you are not correct, because there was no previous representation that h. pylori bacteria caused ulcers, until the Australians created it. Because of this mistake, you have missed the point the GP was making: the arbitrar of what is a true and what is false comes down to a human judgement, and Google is likely to base that on some type of concensus, since basing it on anything else would be problematic. In other words, if Google's plan is to say, "Well, everyone thinks that vaccines don't cause autism, but we know they do." and rank pages accordingly, that is a problem. If they go with consensus in every case, then there is a bias against corrective information. The consensus now is that h. pylori causes many ulcers. That is a representation... a fact. Is it a true fact, or a false fact? How do you know? Are we really any different than the folks in the 80s, who "knew" ulcers couldn't be caused by bacteria? Because we don't have any false facts now? That we know of?

it's slashdot!

[lkcl]

yaay, it's slashdot on steroids, yaay! with all the means by which true knowledge may be suppressed by misunderstandings, yaay! democracy at work to bubble up the sum of our ignorance rather than inconvenient and annoying truth. ahh gotta love it...

Google is becoming useless

[Ozoner]

This sounds like a great improvement.

Some years ago if I searched for a data sheet for an Electronic Component, I could rely on a direct link to the PDF in the first hit or so.

Now however, any worthwhile result is often many pages down the list. The first page or two are full of "Are you searching for xxxx? We don't have that right now, but here's a great way to earn big dollars!!".

Google is so badly scammed that I usually don't bother. I hate to say it, but even Bing is better now.

Re:Google is becoming useless

[ortholattice]

Out of curiosity I recently tried to find out about a product, Power Innovator, whose annoying ads claiming to cut your electric bill by 80% keep popping up on various sites. While it is obviously a scam to anyone with the slightest knowledge of physics, they really have Google fooled. No matter what you type, "Power Innovator review", "Power Innovator scam", "Power Innovator ripoff", etc., every link, page after page, is a "review" or a page questioning "Is Power Innovator a scam?" each ending with a link to buy Power Innovator. I was unable to find any page clearly stating the obvious fact that it is a scam. I feel sorry for all the misinformed people who are sucked into this, and the company must be raking in a fortune of ill-gotten gains. This is a case where Google is completely useless.

Follow the herd or vanish

[Anonymous Coward]

In other words, if your web page contents do not agree to some arbitrary consensus as defined by the pages Google chooses to trawl, your web page will not be listed anywhere near the top of the search results.

This is idiotic, as this has nothing to do with facts, and everything to do with conformance and not rocking the boat.

However, as a business plan, this might actually work: it will be easier to package the products to the advertisers, as all possibly controversial information is removed from the searches.

I for one welcome our Corporate Overlords!

By facts, not links?

[LynnwoodRooster]

Well there goes Wikipedia!

Re:By facts, not links?

[Shadow of Eternity]

It had come a long way, then it started being manipulated by ideology pushing extremists that have become very adept at abusing the hell out of labrynthian policies to the point that even when the author of a news article flat out says "They're lying, I never said that at all" it's the author that gets punished.

What could possibly go wrong?

[Karmashock]

While in theory the idea is great, the problem is that one person's facts are another person's propaganda.

Look at the crap storms on wikipedia for example with all sorts of various groups all fighting over who gets to edit some page. Can you honestly say that always ends with the people standing up for truth winning? I can think of a few situations where it was controversial and the people that were pushing bs just happened to win or nearly as bad force moderators to lock the listing in a pre crisis state. Thus basically white washing the whole incident out of existence.

Again, I think it is a nice idea in theory, in practice I'm sure assholes and trolls are going to fuck it up.

Re:What could possibly go wrong?

[Karmashock]

The simple fact that you see partisan advantage proves you agree it will be exploited. You just think it will be exploited to help you rather than hurt you.

The funny thing you're not getting is that google itself is far from indispensible. There are lots of other search engines and most of them are comparable in quality.

At best... assuming you're right... what would happen is that people that thought more like you would be even further echo chambered then you already are... protected from alternative view points and increasingly controlled by your ruling opinion makers... while the internet culture further fragments.

You're hardly going to win against your political rivals because they'll just use Bing or something. All that will happen is that YOU are aware of less. Will your opinions change? You weren't moved by sites that came up that you felt were pushing disinformation in the first place. No one really is at this point. We can smell our own and we treat with skepticism anyone outside our tribe.

You might argue that this would help you fight for the allegiance of the moderates and the middle. And you might have something there. But things are so social networked at this point that you'll have a hard time playing gatekeeper with any proficiency.

In any case... your opinion has merely validated my initial statement.

Re:What could possibly go wrong?

[Karmashock]

While of course it would be politically convenient if your rival political factions in a democracy were shut out from any ability to coordinate, debate, or express themselves... there are a few problems with your fascist agenda.

First, you are admitting I'm right in that you agree it could or even should be used for propaganda purposes.

Second, in shutting out your rivals you're going to marginalize google in the first place. They'll just move to other systems. Look at Fox News or Rush Limbaugh or the Wallstreet Journal. Do you honestly think you're going to shut them down? They'll just create something else that will rapidly be just as effective. There are a dozen search engines besides Google. If google says "fuck you" to half of America then half of America will just start using Bing or something. So you'll have accomplished absolutely nothing besides google losing half its marketshare to its rivals by pointlessly pissing people off.

Third, your whole line about the "common good" is basically the same old "greater good" argument that the Nazis were fans of. You want your rivals silenced by force rather then meeting them in open public debate. This shows both that you're not actually a fan of democracy, that you're willing to impose your views by force against the wishes of others, and that you'll have that same blind faith in your own actions found in any fanatic while they do it. So... basically you're a terrible person. No offense intended. But the shit you're typing is indefensible. You can't possibly be a liberal because you don't believe in "liberty". What you believe in is the orthodoxy of your dogma and you're prepared to subvert democracy to get what you want. Here is a lesson for you, sport. The means ARE the ends. That is, the process by which you create something has a major influence on what you ultimately create. If you gain power through trickery, extortion, force, etc then your power will not be based on the truth or the will of the people but rather on your trickery, extortion, and free use of force. That will be the nature of the society you create. And that will furthermore justify your opposition to respond in kind. If you lock them out of debate then they will be entirely justified to respond by shutting YOU out of debate or just ceasing power or something.

I don't know what to tell you. I find posts like yours to be disappointing. You're basically saying you want a king or an oligarchy to cease power and rule with an iron fist... managing everything with some digital ministry of truth. You make me sad.

FEO

[Kunedog]

If this ever happens, expect Fact Engine Optimization to become a new industry, and do exactly what SEO did to the reliability and utility of search engines.

Re:FEO

[Anonymous Brave Guy]

"Fact optimization" is already behind more than one multi-billion dollar industry: advertising, political lobbying...

And this is why I fear this initiative, no matter how well intentioned, is doomed to failure. Just because something gets repeated a lot, that doesn't make it factually correct. Moreover, censoring dissenting opinions is a terrible reaction to active manipulation and even to old-fashioned gossip, because it removes the best mechanism for correcting the groupthink and promoting more informed debate, which is introducing alternative ideas from someone who knows better or simply has a different (but still reasonable) point of view.

Remember, not so long ago, the almost-universal opinion would have been that the world was flat.

Re:FEO

[Mr. Shotgun]

If this ever happens, expect Fact Engine Optimization to become a new industry, and do exactly what SEO did to the reliability and utility of search engines.

Finally! It is the tautology club's moment to shine

Re:FEO

[thieh]

It'will be interesting to see whether websites of some particular political party or lobbying group get downranked so much to disappear from search results completely.

Re:FEO

[dryeo]

Remember, not so long ago, the almost-universal opinion would have been that the world was flat.

A good example of a wrong fact that too many people believe. As soon as people really started traveling, especially on the ocean, it became obvious that the Earth is not flat. Something like 2500 years ago a Greek used geometry to measure the circumference of the Earth though there were idiots like Columbus who were convinced the world was much smaller then the generally accepted size.

Re:FEO

[AK Marc]

Yes, 2+2=4, therefore global warming is a lie by Al Gore, who served as Vice President. I'm sure the linguistic style will look odd at first, but packing one lie per page and lots of valid facts could game the system. That's why they should release it early, so we can start gaming it early, so they can improve it before release (but they'll get around it by calling everything a beta).

Re:FEO

[dryeo]

Don't be daft, good vs bad are not facts. Facts are things like when the bank bailout happened, which laws were used, who signed the laws, which banks benefited, how big of bonuses were given out etc. Ideally you base your opinion about good vs bad based on facts rather then bullshit and good vs bad is always an opinion.
And how is it censorship if a private entity prints whatever it wants? You, I and Google are free to put whatever we want on sites we own. Everyone is free to visit which ever sites they want to visit and we're all free to stop visiting a site if we don't like/agree with its content. Google fucks up and they'll go the way of Alta Vista.

Re:FEO

[CrimsonAvenger]

though there were idiots like Columbus who were convinced the world was much smaller then the generally accepted size.

There is a certain amount of evidence that Columbus lied about how big he thought the world was, in order to convince the Spanish crown to finance his expedition.

It's not like the New World was completely unknown in Europe before Columbus - FLemish fishermen were drying fish in Newfoundland before Columbus was born. And it's quite possible that Columbus knew that.

If so, and in light of Spain's interest in breaking the Portugese monopoly on trade with the Far East, a little "creative interpretation" of the world's size might have been sufficient to convince the Spanish Crown that a trip west was a worthwhile investment....

Re:Search Neutrality?

[TapeCutter]

Maybe there should be a concept such as Search Neutrality.

Search is already neutral, it's the order of results that upsets people because by definition, ranking cannot be neutral. An intelligent search engine such as IBM's Watson gives a confidence rating on it's "facts", it can produce multiple answers each with a confidence ranking. They can also explain in excruciating detail how they arrived at the answer, and they can do it better than humans. Technology is a tool, a hammer that can be used to build or destroy a civilization.

This shifts the weakness in Google's rankings

[aussersterne]

from gameability (in short, SPAM) to politics. Rather than punish above-board or non-predatory websites, it will punish both subversive and innovative thought that runs well ahead of social consensus. Sure, it will also eliminate willful misinformation, but it turns Google into an inherently conservative, rather than socially innovative, force.

Can't say I think it's better. Probably not any worse, but certainly not panacea.

Re:Sell any stock before they launch this...

[sumdumass]

You are not envisioning a fact based result but a "your opinion" based result. Not really what is discussed here. Fox News for instance, gets more facts right then wrong even though they are selected to shill for the republicans. You have no facts stating that _ALL_religious_websites are wrong.

Who decides as to what is correct

[EmperorOfCanada]

If a site says the world is flat and is filled with rationals about it being flat then that site should probably come up on a "flat earth" search. Yet we can all agree that the facts in this case are completely bogus.

But what about inconvenient facts, for instance the various western governments put out employment numbers that are pretty hard core "facts" yet other people will look at the same "facts" and realize that they have had massive amounts of spin put on them. For instance in my neck of the woods they desperately hide the fact that most jobs being created are really crappy. Thus these "facts" then become politicized.

Or what about someone writing about NSA evildoing? Those are facts that the government would love to go away. Or what if every stock analyst suddenly agreed that Google was doomed as a stock?

Then there is group think. Prior to the 2008 financial crisis there were some "crackpots" who called it exactly and made fortunes based on their predictions; yet those facts flew in the face of general consensus. The same in economics. One joke at many economics universities is that the questions never change on the final exam, it is the answers that change year to year. If you look at something such as to the best time to loosen monetary policy and every major economic school has its own "facts".

I don't think that Google's search engine problems come from facts it is more that SEO whores like huffpo or the various directories are driving all the results to their crap sites. I don't know how many times I have searched for a company that has a perfectly good site that has not been through an SEO pimping putting it on page 3 or more while the first many pages are all kinds of crap yellowpages that ask "Is this your site?" where they want to upsell the owners on crap services.

Re:Sell any stock before they launch this...

[Oligonicella]

All of which will still outrank CNN. :)

And Google's epistemological basis is?

[matbury]

I wonder what the Google staff's and consultants' philosophy of epistemology is. What do they mean when they say fact? What assumptions underly that definition? Are they naive positivists or social constructivists? Ultimately, it requires people to decide what constitutes truth, fact, and knowledge - machines are nowhere near being able to do that and perhaps never will. Do they expect to automate this ranking system with an algorithm? I can't wait to see it trip up over criteria-matching random string generators that regurgitate scraped "facts" off the web (by simply following Google's own "fact" ranking results) to push their porn, malware, and sales/scam/phishing sites up to the top of Google's page rankings.

This post was brought to you by Carls Junior, makers of Brawndo, the thirst mutilator. It's got electrolytes.

My personal metric

[justthinkit]

Good choice.

My personal metric will be any web page that doesn't contain the string "silentcoder".

Re:My personal metric

[Neil Boekend]

You know you can append your search with "-silentcoder" to exclude all results that contain the string "silentcoder", right?

Re:Wonder how they'll rate Global Warming discussi

[Ungrounded Lightning]

In particular, I wonder how they'll handle Global Warming / Climate Change discussions.

That's a simple one. There's not really a dispute there.

Q.E.D.