A couple of points based on reading Japanese news reports:
The reports are vague (at least to me) as to whether the injunction covers autocomplete itself or just the particular result. They say Google was ordered to stop the 表示 (hyouji, lit. 'display'), which I think could mean either, but seems more likely to mean the particular result (otherwise I suspect they would have said 'function').
In any case, it should be noted that all the reports seem to be based on a press conference by the plaintiff's lawyer, not on any court documents. Also, the English reports are reporting it as an injunction, but the Japanese sources have 仮処分 (kari-shobun, an interlocutory injunction), i.e. this is a temporary order and the case itself is probably still ongoing.
(I am not a native Japanese speaker, but professional J>E translator with legal translation experience)
"Google has responded to the man's complaints by saying that since the results are compiled automatically there is no intrusion of privacy"
It's hard to see how this case is about privacy - it seems like it's really about defamation, however the real details aren't disclosed.
That said, the argument that because something is done by a machine the operators are not responsible for the consequences needs to be debunked once and for all.
Google has so far refused to take action, saying Japanese law does not apply to its US headquarters and its own corporate privacy policy, Tomita told reporters.
I found this quite funny given all the recent events regarding domain names, MegaUpload, PirateBay, MAFIAA etc etc. Now if only the US would take the same view and leave the rest of the world alone...
Up until you said that, I assumed that was the general feeling. Though, I've never been in this situation myself. Can you qualify why that attitude is disingenuous and unhelpful?
Disingenuous: Because if you genuinely can't, you don't control your algorithm, you're just there to keep the lights on. I refer you to the joke about the autopilot computer, the pilot, and the dog.
Unhelpful: because it is. This is someone who is being deeply distressed by Google's actions in this small area of their operations, and their reaction is to tell him to take a hike. They don't have any obligation to him, but it's not good corporate citizenship. Vaguely related rant incoming; have some patience, please. (",)
We make algorithms, feed them masses of data, and have them suggest things to us based on our previous measured behaviour. This produces a positive feedback effect; in news and current affairs, you gradually get directed more and more towards news confirming your biases (see the http://en.wikipedia.org/wiki/Filter_bubble concept).
So naive implementations of news suggesters should be "considered harmful" to the public discourse. But those implementing these things don't seem to care, and often seem to show exactly the "it's the algorithm, not me" attitude displayed in this situation. We have a greater responsibility than that; if we're going to cause the (almost-)elimination of the job of newspaper editor, I'd hope we'd do it with something capable of doing at least as good a job.
"if we're going to cause the (almost-)elimination of the job of newspaper editor, I'd hope we'd do it with something capable of doing at least as good a job."
That's a fair request. Though, google didn't design search or autocomplete to do that right? Are they held responsible because some people figured it was good enough to do that job? I completely understand that it is detrimental to some people. At the same time, if it is changed, doesn't that take away its core functionality, which is actually useful to everybody?
Search and autocomplete serve a specific purpose: to search the internet for topics relevant to a query, and to aid in searching for common topics, respectively. People give them more functionality then that. Is that google's fault? Would they be fair in saying, "You're using it wrong"?
I did put "vaguely related" in the rant warning, didn't I? :-)
Though news.google.com is what I was thinking of with the current affairs example, it generalizes to the whole Google edifice, and to Facebook too. I really suggest having a look at this: http://www.youtube.com/watch?v=B8ofWFx525s to see what I'm talking about, with actual examples.
Minor quibble: autocomplete isn't to aid in searching common topics, it's to help Google avoid failed searches through misspelt queries. ",)
Because algorithms aren't arbitrary physical laws, they're created deliberately by people.
"My spambot sends spam because the algorithms it follows tells it to, so it's out of my hands," wouldn't fly, and I don't think it's an appropriate defense here, either.
Why is direct mail marketing unethical, but ignoring the very real distress linking this man's name to suggested criminal history isn't? We have to consider the effects what we do has. When we only had local impact, if we did something silly or damaging we could guess the impact, and quite possibly see everyone affected. In these Internetted times, if a product can do something silly or damaging it may have spread around the globe before it's found out. We need to be aware of this.
(Note: not debating the morality of spammers; merely asking a Socratic question.)
And jlarocco was only using spam as an example of an algorithm; telling someone "I can't turn my spam-sending program off" is as silly as saying "I can't stop that autocomplete result showing." It's a difference of magnitude, not of kind.
Auto-complete is just a subset of search. Why is it more offensive to type "john smith" and get the suggestion "serial killer" than search "john smith" and have the first three results being about John Smith, the serial killer, and only getting to John Smith, the mountaineer and cat lover, in position four?
I think that one valid argument could be if people searching for this guy were commonly using a term like "John Smith Gym Teacher" they would likely never see the results for "John Smith Serial Playground Rapist". Or, at the very least the playground rapist results would be so far down the list that they would naturally be mostly disregarded. But, in this case the fact that "Playground Rapist" results are more common than the "Gym Teacher" results it could be argued that this skews the perception of John Smith in peoples minds.
I'll go against the popular consensus and argue that Google needs to turn this feature OFF by default. When ON, it's far too easy to assume the auto-complete represents the "right" search terms. IMO, it should be clearer that what it actually represents is similar searches other people have performed.
My point is that auto-complete makes unwanted and unnecessary associations between terms. I have no problem with "Did you mean ... ?" suggestions appearing after I've searched for something; I simply disagree with Google "talking me into" search terms I didn't explicitly ask for.
Did Google even defend against this in the Japanese court? I don't understand how this could get by if they did.
The fact is that names are not unique. Would he be complaining if his namesake was a renowned scientist / businessman / whatever?
Just because someone with his name is linked to crime does not mean that he is. It simply means that someone with a similar or same name is known for their criminality. I don't have what I would term as a popular name but hundreds of people share it. I would imagine in countries with far denser populations that meeting someone with the same name would be fairly common.
I would also like to see an employer who received a CV. Went to Google and typed in half a name and without even searching discarded the CV because auto-complete had shown the name along with a buzz word like crime.
Your name may not be unique. That's not necessarily true for everyone else.
Given the prevalence of search, I would not be surprised at all if employers (more specifically recuriters) did a search and discard. It does not mean it's right or ethnical. But afaik there are no legal ramifications currently. Which is unfortunate.
Those auto-completes can be gamed, as most of them are just searches most people have done. What you do is get a large group of people to search for <NAME> + something else.
Of course you need to be creative, so rather than tell your friends, make it a contest or something. Or in your Linked In profile and email signature, put something like "If you want to see my latest resume, just google "<NAME> resume".
Google bombing usually means artificially trying to get a webpage to appear #1 for a specific keyword. This technique is not about trying to rank, but trying to get a phrase to appear in the search suggestions. Not many SEO blogs have even touched on this...
I have found one thing disturbing about autocomplete: Suppose I find a new "association" between 2 ideas, one of which is rare. I intend to follow the relationship up as a researcher or a programmer. Because of the rarity of the first topic, it is suggested to anyone who types in the first keyword - (s)he discovers the association for free, courtesy of my flash of inspiration and my carelessness to search it on google.
It is surprising in the history of mathematics how often the same results are proven simultaneously in several places - see for example, the HOMFLY polynomial. It is as though the "result was in the air". Now this is a concrete way of leaking such information so that if any one makes a connection, it "leaks" to others.
Dear sir, I would like to apply for the position of squirrel hairdresser. Please note, I am not the Bill Godfrey recently convicted of lightbulb smuggling. That's someone else with the same name.
I thought people also have the ability to turn it off themselves - at least i thought they did, I'm struggling to find the option now. There is also another option to turn off instant update of the search results, which annoys the hell out of me (this is only for people logged into their Google account)
In any case, I'm sure it's just a case of Google flipping a switch somewhere...
I wasn't planning on it, but I feel somewhat spoiled by your warning, actually. Maybe I shouldn't, but I can't help make certain assumptions when given a character's name and that there are major spoilers on the level you've described. : /
seems like that is what they are going for (according to the text of the article, not the title):
"...financial damages in a bid to press Google to erase the suggested search..."
(I am not a native Japanese speaker, but professional J>E translator with legal translation experience)