(34)
@Skitzzo shared a photo of a search he did on Google with me on Twitter showing just how vulnerable Google's Knowledge Graph is.
A search for the [St Louis Cardinals] yesterday afternoon brought up information in Google's Knowledge Graph sourced from Wikipedia. But Google indexed the Wikipedia content during the short period that the content was vandalized. For a few hours last night, Google showed the vandalized text while social media was sharing and mocking Google's search results.
Here is a picture:
This shows how potentially weak the knowledge graph can be. Especially during seasonal topic times, such as the World Series. A Boston Red Sox fan needed a computer and good timing to pull this off.
We know the knowledge graph is far from perfect. We've seen nude images, wrong marriages, and wrong pictures to name a few in the past. Here is one more to lighten up your day.
Forum discussion at Twitter.
Comments:
Rob Kissell
10/29/2013 01:43 pm
In school we were never allowed to cite Wiki for ANYTHING... maybe Google should take notes...
Bill Nadraszky
10/29/2013 01:54 pm
Actually my kids are in middle school now and still not allowed to source wikipedia. Crowdsourcing is not always accurate as we see above
CR
10/29/2013 02:13 pm
seems legit...
Michael Martinez
10/29/2013 02:17 pm
Google will probably never stop citing Wikipedia as a source of good content. It's one of their huge failings in intelligence that is completely inexplicable.
Roie S
10/29/2013 02:28 pm
The dangers of scraping other sources content
Rob Kissell
10/29/2013 02:30 pm
Sacrificing quality for the sake of speed of growth in features...
LLBDub
10/29/2013 03:02 pm
isn't that an accurate description of the Cardinals ?
Durant Imboden
10/29/2013 03:53 pm
Garbage in, garbage out.
JustConsumer
10/29/2013 04:00 pm
We will see more and more of the evidences, that Googlers didn't do even a small part of the innovation, related to search, they are proudly talking about. Here is my input : http://www.wtff.com/p.php?h=Google_23
auntyem
10/29/2013 04:43 pm
fyi: nsfw if you check out the rest of the site
Josh
10/29/2013 05:03 pm
Waits for Googlepedia...
JustConsumer
10/29/2013 05:04 pm
Run away from such job or ask for higher salary at least )
xoxo
10/29/2013 05:29 pm
coming soon googlecrappypedia. all wikipedia doorway pages, all amazon products not available pages, all spammy youtube videos, all porn articles (just try to search google for porn and you will find in-depth-articles(!!!) and other nice spammy things from google.
xoxo
10/29/2013 05:30 pm
google not care about search, they want only bloody $$$$$.
Durant Imboden
10/29/2013 09:04 pm
IMHO, this silly stunt says more about Wikipedia than it does about Google. If someone hacked into the National Weather Service's site and fudged the forecasts, would Google get the blame for inaccurate data from the NWS on its SERPs? As I said below, "garbage in, garbage out." Wikipedia mostly does a pretty good job, but it would be a better resource (and source) if its edits had to be approved before going on public display.
JustConsumer
10/29/2013 11:38 pm
Sure, it's about the Google. Google shows wrong information, pretending to be the smartest search engine. This is an algo problem, if algo doesn't consider such occasions. "its edits had to be approved before going on public display." Exactly same can be said about Google as well )
Alan
10/29/2013 11:38 pm
OMG Durant seriously can Google ever do any wrong in your eyes? We get it you love Google! we get it! This is not a conversation about how bad wikipedia is (we all know how bad wikipedia's data is, they do not claim to be an authority on any topic, they just claim to be a repository that is user sourced.) This is a conversation about Google accepting the data at wikipedia as gospel.
xoxo
10/29/2013 11:53 pm
google is evil! they know everything, they spy everybody but still unable to see what happens in residue of their serp.
Durant Imboden
10/30/2013 01:04 am
First, ditch the ad hominem argument. People who attack the messenger instead of the message usually do so because they have no better argument to make. Second, let's recognize this incident for what it is: An adolescent prank that's mildly amusing to frat boys. It isn't the 50-caliber smoking gun that some people would like to think it is. Third, the snippets that Google uses for "Knowledge Graph" entries on its SERPs are just snippets. Is anyone really stupid enough to think the St. Louis Cardinals are a "gay butt sex team" because someone vandalized the source of a snippet? Fourth, Google Search results aren't perfect. (Neither are Google News results, for that matter: I see outdated or incorrectly-geotargeted headlines on Google News all the time.) Algorithms are dumber than people. Sometimes they're even dumber than adolescent pranksters.
Alan
10/30/2013 01:14 am
Yes but you said this says more about Wikipedia than it does Google and Durant right there you are wrong. Simple as that. Wikipedia has never claimed to have un-tainted results. Infact they openly admit that their results are open to abuse. So Google should not be using them for knowledge graph. In fact knowledge graph sucks anyway but that is another argument.
Durant Imboden
10/30/2013 02:16 am
Again, garbage in, garbage out. Most of the time, snippets taken from Wikipedia are just fine. (And remember, we're talking about snippets here: simple answers to questions like "capital of illinois" or "birth date of mother teresa.")
Alan
10/30/2013 02:53 am
This comment made my day :)
PM Fiorini
10/30/2013 03:25 am
Where is Michael Martinez when you need him
JustConsumer
10/30/2013 03:25 am
Omg, Durant this is IT !!!!!! Not literature discussion club. You don't understand what is the algo and how the algo works. The main question any algo has to reply is : What to do IF ... ? This case shows, that algo failed to reply this question properly. Event IF happened, but algo didn't react. This is the hole in algo. It's unreliable. That's it. Ok, probably it will be hard for you to understand this. Here is an example from the real life. I've been in AnyCity and searched for the restaurant. Google search provided results. I selected. Went there and ups ... it was out of business. Who was responsible, that I lost time ? Was it restaurant owner ? Sure not. Was it me ? Sure not. Google provides information and should be responsible for it. Don't provide, if not sure. Will you publish article on your website, if you are not sure about the information? I believe not. You'll ask yourself the question "What will happen IF ... ?" Why you, me, most of the people are responsible about the information published and Google is not. Just because it's Big ? How is this an excuse? Do I need to remind you, that information is the core of the modern society ? Do we need to wait, until something really bad happens because of false information, provided by Google ? "Big" means not only big money, but also big responsibility. Your take on this is just immature. P.S. By the way, did Google apologize at least ? Rhetorical question.
Craig Hamilton-Parker
10/30/2013 08:42 am
Perhaps they should use Encarta?
xoxo
10/30/2013 09:06 am
high authority truth :) even here - google cannot trust anybody. So looks like no space for google in the web to trust/believe anybody. They already dropped all small/med sites, interesting to see what they will do about wikipedia false information & spammy pages. pandalize & penguinize it? (I think not, because in this case no will results to display. It also mean what rules defined by google is not for everybody, only for small guys). Even schools & academic organizations not rely on wikipedia datas, but google knowledge-base based on it. Just why it called not google wikipedia knowledge graph???
xoxo
10/30/2013 09:11 am
so google search now = wikipedia. It funny what they show scrapped information as own (on google pages). Why they not change name to wikipedia search engine & graph?
xoxo
10/30/2013 09:18 am
it not "smart" search engine anymore. The smartest google was in pre-2012, when you was able to find answer on your question (you was able to REALLY search the web). Now google operates with VERY LIMITED set of datas, so they just unable to provide us quality results. This is all about $$$ and adwords conversions, not about user experience
Bhatt Hardik H.
10/30/2013 11:08 am
here what i found about google penguin 2.1 recovery : http://google-penguin-penalty-recovery.blogspot.in/2013/10/recovery-from-google-penguin-21.html
Amit Kumar
10/30/2013 12:55 pm
Another desperation to earn $$$$$ see the ad is this really related to industry or google keeping advertiser in darkness?
xoxo
10/30/2013 01:33 pm
ops... it already here...
Graciousstore
10/31/2013 02:53 am
The knowledge graph is still in it very early age, it is not surprising that there are many flaws,. Google has plenty of work to do in fixing all the errors.
Yo Mamma
10/31/2013 12:15 pm
Too funny. This just proves that not everything on the internet is real. Hackers rule
Brian D
11/04/2013 11:01 pm
wot?