Is Google's Knowledge Graph Good or Evil?
Google recently introduced Knowledge Graph, touting it as the next step towards semantic search, a more intelligent search that knows what you’re searching for, and understands relationships between people, places and things. It appears to be a useful time-saver for certain popular search topics. But in its initial debut, some observers are seeing a threat to the fabric of the web...
What Is Google's Knowledge Graph?
The Knowledge Graph is a dazzlingly ambitious project. Google is attempting to map – or "graph" – every thing in existence - people, places, objects, and ideas. So far, the Knowledge Graph database contains over half a billion "things," to use Google's own technical term. But that's only a start; Knowledge Graph also maps the relationships between things, for an additional 3.5 billion database records. Even so, the Knowledge Graph includes only a fraction of all the things and relationships in the universe, and will never be complete. But it does some interesting things for search.
In Google's announcement of the Knowledge Graph project, they highlight the search query, "taj mahal" as an example. Those two words may refer to multiple things: a building in India, a musical group, or a Donald Trump resort, for starters. A traditional search might yield a page of results that mix up all of these very different things; the user must pick through irrelevant results to find the one(s) desired. These traditional search results are still displayed by Google, but over in the right-hand sidebar is a new Knowledge Graph result.
When a search term is ambiguous, Knowledge Graph presents multiple options and asks you to choose the set of results that you really want. Choose the building in India, and Knowledge Graph displays a set of links to things and search results that pertain to it. Without clicking through to a new site, you can learn where the Taj Mahal is, how tall it is, when it was built, its architect's name, even its phone number. You also get a brief summary of what it is, credited to Wikipedia. But you never have to leave the Google search results page.
This is all good for users who just want a quick summary of a subject. It's also good for Google, which has an interest in keeping users on its pages instead of sending them off to Wikipedia. But what about Wikipedia, and other sites that depend on visitors for their livelihoods?
Google has an inconceivably huge stockpile of "things" that it has acquired from other sites on the Web. Is it fair for Google to "borrow" the facts and figures compiled by others, and present them prominently on their search page? By combining things that it has essentially appropriated, Google can, conceivably, become the only site that users need to visit in order to learn what they want to know. That could pose traffic problems for other sites, and raise antitrust questions from regulators. It also looks quite similar to the automatically produced content created by "scrapers" that Google disdains and punishes.
Looking Further at Knowledge Graph Results
Not satisfied with Google's hand-picked Taj Majal example, I decided to see if Google could help me quickly answer other questions. So how old is Mick Jagger, anyway? As soon as I typed "mick j" in the search box, my answer appeared in a huge knowledge graph to the right of the regular search results. Okay... he's 68, married twice, and has 7 children. That could be enough to stop some searchers from visiting the top search results in the left column, which include the artist's official website, his Wikipedia entry and an IMDB profile? Could even make a grown man cry.
But what about websites and products that directly compete with Google? My searches for "bing" (the Microsoft search engine) and "iPhone 4s" (a direct competitor to Google's Android smartphones) do not have knowledge graphs displayed alongside the results. The search term "office" does have a knowledge graph, but it's all about NBC's comedy show. There's no mention of Microsoft's hugely popular Office software in sight, except in ads paid for by Microsoft.
Google touts the ability of Knowledge Graph to help you search for people, places and things. But other terms that don't have associated knowledge graphs include "amazon", "facebook founder", "jesus christ", "nascar", "obama care", and "ralph kramden".
I doubt that this indicates any bias or conspiracy. Right now, the Knowledge Graph is still under development, and results will satisfy only the sketchiest curiosity. Most users will click through to original-content sites to find more in-depth information. But where will Google draw the line between being a directory service and a content producer?
And if they're heading in the latter direction, to what end? Google wants people to click on ads. If searchers visit a content site, and click on a Google ad, Google has to share a slice of the revenue with the content producer. If searchers click ads on a Google-owned page, of course Google gets the whole pie. But in all the Knowledge Graph examples I've seen, there are less ads on the page, and none within the Google-supplied facts and figures.
If people are finding answers on Google, pasting them into their word processor, and leaving without clicking any ads, Google loses. And if content sites die off due to lack of traffic and/or revenue, everyone loses.
There is an implicit contract between Google and the sites it indexes. The sites give Google permission to index their content (and even display snippets of it in search results). Any site may deny such permission by including a "do not index" tag in its HTML. Most sites are happy to be indexed by Google in exchange for visitors that Google sends to them via search. But is Google's Knowledge Graph breaking the compact with content producers, by cutting them out of the loop?
Your thoughts are welcome on this topic. Post your comment or question below...
This article was posted by Bob Rankin on 24 May 2012
|For Fun: Buy Bob a Snickers.|
Free Windows 7 Repair and Recovery Tools
The Top Twenty
Wolfram Alpha - Join the 0.1 Percent!
Post your Comments, Questions or Suggestions
Free Tech Support -- Ask Bob Rankin
Subscribe to AskBobRankin Updates: Free Newsletter
Copyright © 2005 - Bob Rankin - All Rights Reserved
Article information: AskBobRankin -- Is Google's Knowledge Graph Good or Evil? (Posted: 24 May 2012)
Copyright © 2005 - Bob Rankin - All Rights Reserved
Most recent comments on "Is Google's Knowledge Graph Good or Evil?"
24 May 2012
I believe the basic concept is great, and I really don't see a downside at this time.
24 May 2012
If you think it's going to be an improvement just try living in the UK and trying to find presents or information for family (including grandchildren) in the USA.
Oddly it's a lot cheaper to get a return ticket from the USA to the UK than it is from the UK to the USA.... BUT try finding out what it's going to cost someone flying from the USA to the UK when you are connected to google in the UK, it doesn't work and we are finding out for our Aunt Denise(71) and maybe Great Aunt Patsy (88).
Anything that takes away the basics of putting in the RIGHT data to get a CORRECT answer doesn't work.
My vote is for EVIL!
24 May 2012
So what you are saying is that this Google Knowledge Graph has Mick Jagger, but not Jesus? Hmmm...perhaps it does have "Sympathy for the Devil!"
In all seriousness, thank you for explaining what this new graph does. I have shared your findings with several co-workers at the school where I work, who help students do internet research.
Knowing that the graph itself can skew a piece of research is important to know.
24 May 2012
As you say "spelling, punctuation, grammar and proper use of UPPER/lower case are important!" That's why I wonder if you did mean "is Google's Knowledge Graph breaking the compact"?
Or was it about a contract? :-)
24 May 2012
Been a while since I used 'Google' so I thought I'd give it a try...
First try for 'mick j', I got the same results you did. I bet Mick Jones is ticked-off!
Second try for 'mick j', I got no 'Knowledge Graph' at all.
Anything Google related brings up the latest entry on 'Google+'.
Interesting, but not enough to bring me back (yet).
24 May 2012
Not the first time someone has tried to graph the world of all known INFORMATION but called it KNOWLEDGE. After all this time I would have thought Google would know the difference between knowledge and information but I guess not. Funny, with all the information at their disosal they don't know the difference between the two.
24 May 2012
In their video introducing this project (http://googleblog.blogspot.com/2012/05/introducing-knowledge-graph-things-not.html) Google puts it this way:
"We're in the early phases of moving from being an information engine to becoming a knowledge engine. And these enhancements are one step in that direction."
That's gotta be good! And it's also a project whose immensity boggles my mind.
25 May 2012
DEAR BOB RANKIN.... AND READERS....
GOOGLE ONCE AGAIN BREAKS THE KNOWLEDGE BARRIER OR SOUND BARRIER.... AND MAINTAINS OR INCREASES ITS EDGE AS THE SMARTEST AND MOST KNOWLEDGEABLE ENTITY ON EARTH....
27 May 2012
"Where is the widom we have lost in knowledge?
Where is the knowledge we have lost in information?" - Thomas Stearns Eliot (1888-1965)
I never believed their "do no evil" mantra in the first place. Think I'll be sticking with DuckDuckGo.
27 May 2012
I saw the founder of Google interviewed the other day on Charlie Rose over PBS. He was giddy over the prospects of their new project and its potential.
When caller-ID first came out, a hardware store answered the phone addressing the caller by name ("Hello, Mr Smith, how can we help you today?").
The callers were turned off because the Hello they got invaded their "space." The hardware store then changed its policy and ignored the caller's name.
Telephone answerers may be told people love to be addressed by name, and one bank asks my first name at the start. To my objection that we're not personal friends, they say, "but then I don't know what to call you." I reply, "then ask."
Google's attempt to read your mind from your search inquiry bothers me. This is an instance of their getting too close, standing, as it were, an inch away from your nose.
The Google founder seems to be a nice guy who is totally oblivious to the implications of his invading our space. And he looks forward to integrating your cell phone and all other electronic input into one portrait.
If this is representative of their thinking, they may do a lot of damage to your privacy without intending to.
Don't forget to log out of Google as soon as you're through your search.
I don't use Chrome because there is no way you can contact a human being for a question their Help tab does not address. Only when I uninstalled it did a field open up asking why. That should tell you a lot.
01 Jun 2015
Maybe it's a generational thing, but I consider that a gross violation of privacy. What if you had to wear a GPS tracker, so somebody knew exactly where you went, how fast you drove, every store you visited, what aisles you went to, what items you stopped to view, every address you went to and how long you stayed? If the government does that, we call it surveillance, and the courts have ruled they need a warrant to even put a GPS tracker on your car, and only certain convicted individuals have to wear a personal tracker; if an individual does that, we call it stalking, and there are legal penalties; if a corporation does that so that they can make more money, we call it business as usual? I'm sorry, I just don't buy it.
The Europeans understand this much better than Americans do - but of course our government has swung almost completely to the side of the corporations over the last few decades; we're no longer "citizens" to them, we're "consumers", who provide 70% of GDP by buying things. And they seem to think that it's fine to inconvenience thousands of people if you can make money at it - e.g flying slow noisy airplanes towing banner ads low over residential areas when there are ball games at Fenway Park. You seem to be playing along with them, by maintaining that gathering and collating highly personal information is harmless, but there's a problem if an extension of that project violate contracts with other businesses and organizations. Say what?
The idea of tying all the world's knowledge together, and making lookups easier, is laudable; in one sense, that's what Wikipedia is about. I don't think, however, that it should be done by a company which is building on the most extensive private database of personal information in existence (although Homeland Security is doing their best to catch up). I'd feel much less queasy about this if it were a consortium-funded project undertaken by Wolfram Alpha or Encyclopedia Britannica, and/or a network of universities, with Google support.