But that hasn’t stopped plenty of people from going after Google when they don’t like what they see appearing next to their names or businesses. And sometimes successfully, I might add. The latest case to spring from a disputed autocomplete result comes from Germany and is bad news for Google.

A German court has ruled that Google must manually remove autocomplete results if they are determined to be defamatory. This wide ruling could have an effect on not only cases in Germany, but in other countries who could use the decision as a model.

As the AP reports, the case stems from an unidentified plaintiff, only known as “R.S,” whose company sells nutritional supplements. R.S. filed a complaint when they saw that Google autocomplete results associated the name of the company with “fraud” and “Scientology” – both of which they considered defamatory.

A lower court dismissed R.S.’ claim, but the Federal Court of Justice overruled. According to the ruling, Google isn’t being directed to turn of autocomplete or even interfere preemptively, only required to eliminate defamatory autocomplete suggestions when they are brought to the company’s attention.

This isn’t the only case in Germany involving Google’s autocomplete to make headlines. Last year, former German First Lady Bettina Wulff claimed that Google destroyed her reputation with their autocomplete suggestions. Wulff, who has battled rumors that she worked as an escort prior to marrying former German president Christian Wulff, has her name associated with “escort” and “prostitute” in multiple languages in Google autocomplete.

Of course, those suggestions only exist because of the high volume of Google searches. But this new ruling could affect that case, which is still pending.

In April, Google lost a case in Japan over their autocomplete function. A man sued Google over suggestions relating to criminal activity – activity he denied. A Japanese court ruled that Google must alter their results and they also issued a 300,000 yen fine (roughly $3,100).

Google has also faced autocomplete complaints in France.

And that’s what makes this so…just…yeah:

“Using billions of searches, Google has prototyped an anonymous profile of its users. This reflects the fears, inquiries, preoccupations, obsessions and fixations of the human being at a certain age and our evolution through life,” says creator Marius B. Well said.

And just FYI, he made this video in a incognito window with no user signed in, no cookies, and with a deleted history. This is pretty much the human condition, visualized via Google autocomplete. Life really is all about sex.

[Marius B via reddit]

Like Google, Yahoo, and most other search engines, Bing offers to autocomplete queries in their search box. While Google calls this feature “autocomplete,” Bing calls it “search suggestions.”

Well, it looks like Bing is suggesting that you search for some pretty disturbing stuff.

I was pointed in the right direction thanks to a reddit post. “Why don’t you have a seat over there, Bing,” it read, referencing Dateline NBC host Chris Hansen’s famous line on the show To Catch a Predator.

Ok, I’ll bite. Here’s what Bing’s search suggestions suggest:

Say what? It appears that Bing is suggesting that I search for “sex games for kids,” and “sex games for kids in bed” and “sex games online for children.” Hm, ok then.

Digging a little deeper with the questionable queries produced similar results. For instance, here’s what Bing suggests when you search for “sex kids”:

And here’s a Bing search for “sex child…”:

Even worse, here are Bing’s suggestions for what I’m sure is one of their (and any search engine’s) most popular single-word queries, “sex”:

Damnit, Bing.

Also, changing your SafeSearch settings to strict does nothing to eliminate these results. When you think about it, it would probably look even worse for Bing if it did, because that would indicate that Bing felt that a search suggestion like “sex games for kids in bed” was an appropriate suggestion for a moderate level SafeSearch.

“Still seeing inappropriate content? SafeSearch uses advanced technology to filter adult content, but it won’t catch everything. If SafeSearch is set to Strict or Moderate and you’re seeing adult content, tell us about it so we can filter it in the future,” says Bing.

But this isn’t a SafeSearch problem, this is a search suggestions problem. You can turn search suggestions off in your settings, but by default they are on. That means that the average person that pulls up bing.com and searches for “sex g…” sees these questionable suggestions.

Now, I guess the next big question is whether or not Bing has a responsibility to filter out these search suggestions.

On one hand you could make the argument that Bing doesn’t have to manually edit which search suggestions it gives for particular queries. The suggestions are clearly based upon popular and recent searches from the Bing community – and if that’s what they’re searching for then hey – let it be.

On the other hand, Google limits its autocomplete results. Here’s what you’ll see when you search “sex games” on Google:

And here’s what you see when you search “sex kid”:

As you know, Google also censors other questionable searches. They won’t give you suggestions for sexual terms like “boobs” or “pussy,” and they won’t even display curse words like “fuck” or “shit” in autocomplete results.

They also censor any search that has to do with the illegal downloading of copyright protected content. For instance, “game of thrones torrent” won’t autocomplete.

Over on Bing, it’s a totally different story:

Bing doesn’t really filter any of the types of searches that Google does. Last year, we pointed out that Bing was suggesting painless ways to kill yourself while Google was displaying the suicide prevention hotline.So, if they’re going with a true hands-off approach to any sort of search suggestion censoring, what’s different about queries about sex games for kids?

Well, it’s the “c’mon, dude” argument I guess. As in, Bing…c’mon dude. It doesn’t help that instead of “autocomplete,” Bing’s version of the technology is called “search suggestions.” So, when you think about it, Bing is suggesting that you search for “sex kids movies” and “sex games with kids in bed.”

C’mon, dude.

I’ve reached out to Bing for comment and will update when I hear back.

UPDATE:

As you know, Facebook partners with Bing for their search results. And you can find the same questionable suggestions inside Graph Search results:

UPDATE 2: A Microsoft spokesperson has given me this:

“We’re reviewing the guidelines for search suggestions related to this type of query.”

Well have to see if anything changes.

This case began in March of 2012 when a Japanese court demanded that Google delete certain search terms inside their autocomplete function – ones that related to a specific man whose identity is still being witheld. The man claimed that when his name was searched, suggestions popped up linking him to criminal activity of which he was innocent. Clicking through to the links provided led user to websites filled with further defamation.

Not only did the plaintiff allege that Google’s autocomplete results caused him pain and personal anguish, but they also contributed to him losing his job and being unable to procure another.

Now, the court has ruled that Google must alter their results in the case of this anonymous man. They also ordered that Google pay 300,000 yen ($3,100) for the man’s pain and suffering – but not the job loss as he couldn’t prove that the two were definitely linked.

Well, it’s another day, another foreign court making a ruling on Google autocomplete. We’ve seen plenty of this in the past. Back in January of 2012, Google chose to pay a fine issued by a French court over the company’s autocomplete results. A local insurance company complained that Google autocomplete associated their name with the term “esroc,” roughly translating to mean “crook” or “swindler.”

Later in the year, Google made another deal in a French case, this time involving autocomplete results that labeled certain high-profile celebrities and politicians as “Jewish.” The complaint was originally filed by French anti-racism groups.

Google has also been in trouble in Germany and Italy over their autocomplete results.

Of course, Google’s autocomplete results stem from an algorithm that is based on prior searches. Google does not manually select which terms pop up when you type in any query.

“Autocomplete is a feature of Google search that offers predicted searches to help you more quickly find what you’re looking for. These searches are produced by a number of factors including the popularity of search terms. Google does not determine these terms manually–all of the queries shown in Autocomplete have been typed previously by other Google users,” says Google.

But that hasn’t stopped courts from ordering that Google manually intervene in certain circumstances.

Since Google Search isn’t rooted in Japan, Google isn’t required to follow this ruling – just like they weren’t required to follow the previous injunction the court issued in the case (and they didn’t). The ruling, however, can be appealed.

Starting today, Google begins the global rollout of new autocomplete predictions for your past Gmail searches.

“If you’ve searched your email for ‘supercalifragilisticexpialidocious’ or other lengthy phrases, it just got easier to find what you’re looking for. Autocomplete predictions in Gmail may now include your past Gmail searches,” says Google.

Also rolling out – contact thumbnails in Gmail search:

Google says that the global rollout of both of these features will take a few days and it even includes Google Apps for Business customers.

We think that these updates will probably go over a little better than the last update to hit Gmail. Last week, Google began pushing the new compose box to all users and there was a audible groan from a good portion of the Gmail-usiing population.

Bing refers to its latest development as “Autosuggest Ghosting”.

“Autosuggest algorithms are able to determine just how likely it is that you want the #1 suggestion with various degrees of confidence,” explains Dan Marantz, Senior Program Manager Lead, Bing Experiences and Query Formulation Team. “This confidence is highest in the two major patterns: Navigation and Search History. Ghosting is a way to pre-populate the query most likely to be used in the search box (blue selected-text style below) in an effort to speed up the time it takes to express your intent and get to your destination. This has seen to help users speed up by over 16%.”

“The design challenge was to focus on simplicity and intuitiveness. The interaction should feel natural and instinctive when you need it, and easy to work around when you don’t want it,” he says. “The simplest solution is to grey-in (or “ghost”) the high-confidence suggestive text and hope you notice. The problem then becomes – how do you accept the suggestion vs ignore it?”

Naturally, he takes a dig at Google.

“Google’s model complicates this by not being clear about what happens when you hit <enter> to submit the query,” he says. “Will the search be for ‘bed’ or ‘bed bath and beyond’? Turns out the query is only ‘bed’ and you need to press <tab> or <down> to select the full query.”

The Bing philosophy, he says, is not grounded in applying already-learned interaction models.

Users can press Enter to accept a suggestion, continue to type through it with something else, or press Delete/Backspace to remove the suggested text.

Those questions are at the heart of a Australian case that just tipped against Google, and are likely at the heart of many cases to come. An Australian high court has found Google liable for libelous content tying a man to organized crime. Of course, Google didn’t create the article that made the references, it simply provided a link to it within its search results.

The man’s name is Milorad Trkulja, and he claimed that Google defamed him by associating his name and image with (untrue) claims of ties to organized crime, both in regular search results and in Google Image search. The jury in the case found Google guilty and therefore responsible for the content that they link to. They’ve been fined $200,000, but are in the process of appealing the ruling (as you would expect).

Is Google responsible for the content that is found using their search engine? Or is this a ridiculous claim to make? Let us know in the comments.

Here’s what the Judge in the case had to say:

The question of whether or not Google Inc was a publisher is a matter of mixed fact and law. In my view, it was open to the jury to find the facts in this proceeding in such a way as to entitle the jury to conclude that Google Inc was a publisher even before it had any notice from anybody acting on behalf of the plaintiff. The jury were entitled to conclude that Google Inc intended to publish the material that its automated systems produced, because that was what they were designed to do upon a search request being typed into one of Google Inc’s search products. In that sense, Google Inc is like the newsagent that sells a newspaper containing a defamatory article. While there might be no specific intention to publish defamatory material, there is a relevant intention by the newsagent to publish the newspaper for the purposes of the law of defamation.

Basically, Google may not want to publish it, but they are publishing the publishers. And since Google’s algorithms are tooled to find said content, they are responsible. Or at least it is plausible that a jury could see it that way. The Judge is clearly unconvinced that this stance is set in stone.

The Judge also differentiated search results pages from Google Image searches. The plaintiff also complained of images tying him to crime figures. The Judge notes that a Google Image search is a more-sophisticated version of cut-and-paste from magazines, and importantly a Google-created page:

As was pointed out by counsel for the plaintiff in his address to the jury, the first page of the images matter (containing the photographs I have referred to and each named “Michael Trkulja” and each with a caption “melbournecrime”) was a page not published by any person other than Google Inc. It was a page of Google Inc’s creation – put together as a result of the Google Inc search engine working as it was intended to work by those who wrote the relevant computer programs. It was a cut and paste creation (if somewhat more sophisticated than one involving cutting word or phrases from a newspaper and gluing them onto a piece of paper). If Google Inc’s submission was to be accepted then, while this page might on one view be the natural and probable consequence of the material published on the source page from which it is derived, there would be no actual original publisher of this page.

You can see just how much of a charlie-foxtrot this is. Which pages are Google’s creation, and which are simply the “consequence of the material published on the source page from which it is derived?”

The jury concluded that Google was a publisher, and was liable for the defamatory content even if they weren’t notified of it yet. Although Google contended that it doesn’t matter if they were notified of the content of not – they’re not responsible – the Judge rejected that notion as well.

It follows that, in my view, it was open to the jury to conclude that Google Inc was a publisher – even if it did not have notice of the content of the material about which complaint was made. Google Inc’s submission to the contrary must be rejected. However, Google Inc goes further and asserts that even with notice, it is not capable of being liable as a publisher “because no proper inference about Google Inc adopting or accepting responsibility complained of can ever be drawn from Google Inc’s conduct in operating a search engine”.

This submission must also be rejected. The question is whether, after relevant notice, the failure of an entity with the power to stop publication and which fails to stop publication after a reasonable time, is capable of leading to an inference that that entity consents to the publication. Such an inference is clearly capable of being drawn in the right circumstances (including the circumstances of this case). Further, if that inference is drawn then the trier of fact is entitled (but not bound) to conclude that the relevant entity is a publisher.[42] Google Inc’s submission on this issue must be rejected for a number of reasons, the least of which is that it understates the ways in which a person may be held liable as a publisher.

Of course, $200,000 to Google is basically nothing. The appeal really has nothing to do with the monetary damages. Google knows that this kind of decision sets an unsettling precedent for their future defenses in similar cases. Google as “automated news agent that’s responsible for what their algorithms pull out of the depths” is a view of Google that the company can’t afford to have stick.

We’ve seen this story play out numerous times over the past couple of years with Google’s autocomplete feature. In August of 2011, Google lost a case in Italy and was forced to remove autocomplete suggestion in its search box that tied a man to the word “truffatore,” meaning con man. A few month later, Google was fined $65,000 because one of its autocomplete suggestions labeled a French man “esroc,” meaning crook.

And this year, Google made an out-of-court settlement with French anti-discrimination groups over a “Jewish” autocomplete suggestion.

Google’s argument in these cases is similar to the argument in the Australian case. We’re not suggesting anything. We’re not defaming anyone. Google’s autocomplete suggestions are based on popularity of terms. That means that if anything, Google users are the ones linking people’s names with unsavory terms. Google’s search results are also based on an algorithm. Just ask Rick Santorum about how much responsibility Google claims in what people find using its search engine.

So, is Google a publisher? If not, what are they, exactly? How much responsibility do you think Google has for what people find using their search engine? Tell us what you think in the comments.

This is all part of Google’s goal of getting you to what you’re looking for more quickly, and with less steps, something the search engine has made tremendous strides on over the years.

The following 10 changes deal specifically with autocomplete:

• #83197. [project “Autocomplete”] This launch introduced changes in the way we generate query predictions for Autocomplete.
• essence. [project “Autocomplete”] This change introduced entity predictions in autocomplete. Now Google will predict not just the string of text you might be looking for, but the actual real-world thing. Clarifying text will appear in the drop-down box to help you disambiguate your search.
• #84259. [project “Autocomplete”] This change tweaked the display of real-world entities in autocomplete to reduce repetitiveness. With this change, we don’t show the entity name (displayed to the right of the dash) when it’s fully contained in the query.
• TSSPC. [project “Spelling”] This change used spelling algorithms to improve the relevance of long-tail autocomplete predictions.
• Dot. [project “Autocomplete”] We improved cursor-aware predictions in Chinese, Japanese and Korean languages. Suppose you’re searching for “restaurants” and then decide you want “Italian restaurants.” With cursor-aware predictions, once you put your cursor back to the beginning of the search box and start typing “I,” the prediction system will make predictions for “Italian,” not completions of “Irestaurants.”
• #84288. [project “Autocomplete”] This change made improvements to show more fresh predictions in autocomplete for Korean.
• espd. [project “Autocomplete”] This change provided entities in autocomplete that are more likely to be relevant to the user’s country. See blog post for background.
• #83391. [project “Answers”] This change internationalized and improved the precision of thesymptoms search feature.
• #82876. [project “Autocomplete”] We updated autocomplete predictions when predicted queries share the same last word.
• #80435. [project “Autocomplete”] This change improves autocomplete predictions based on the user’s Web History (for signed-in users).

Last month, Google Autocomplete stopped excluding the term “bisexual,” attracting some headlines for the feature – probably the most positive headlines the feature has seen in recent memory, given that they didn’t involve Google getting in trouble for making controversial suggestions about specific people.

“In some cases we do prioritize the hotline and we’re reviewing the guidelines for instant answers related to this type of query,” says Stefan Weitz, Senior Director, Bing.

ORIGINAL ARTICLE: Earlier this month, Bing launched their “Bing It On” challenge – a blind comparison test designed to see if users preferred the search results from Bing as opposed to those of Google, if they weren’t biased. Bing said that in these blind tests, internet users chose Bing over Google by a 2 to 1 margin. The campaign has some detractors, as some pointed out that Bing was excluding features like Google’s knowledge graph from the challenge results. It could be argued that Knowledge graph is one of the things that people really love about Google nowadays, so that wasn’t exactly fair. And Google’s Matt Cutts pointed out a pretty big fail within Bing’s search results.

Oh well, all of that is beside the point, except to frame the background for this: Bing totally bests Google’s search results, if your search queries involve suicide.

Check out this comparison of Google and Bing results for the search “how to commit suicide,” courtesy of redditor naidlm:

As you can see, Google inserts the contact information for the National Suicide Prevention Hotline at the top, before any search results. Also, Google’s autocomplete will not suggest the full phrase as any point in typing the query. Bing, on the other hand, autocompletes it for you and lists a bunch of related searches on the right.

Additionally, “how to commit suicide” searches on both Yahoo and Ask.com display the number for the National Suicide Prevention Hotline above all results. Neither Yahoo or Ask.com will suggest the complete phrase at any point.

Then, there is the response to other phrases like “how to kill…” Bing autocompletes it with “yourself” and “yourself painlessly” while Google simply suggests bugs.

To be fair, neither Google or Bing can imagine all possible suicide-related queries and plan for them. Searches for “how to slit my wrists” appear similar on both sites. And both sites freely autocomplete phrases about committing various acts of homicide. No “get help before you kill mama” warning from Google.

But searches for the phrases “I want to kill myself,” “I want to die,” “how to die,” “suicide” and “how to commit suicide” bring up the number for the National Suicide Prevention Hotline as a feature in Google search. No such luck with Bing.

We’re not suggesting that Bing should nerf their autocomplete results. Plenty of people take issue with just how many words appear on that Google autocomplete blacklist. But it wouldn’t be too hard to throw a phone number for a hotline at the top of the results when someone searches one of these loaded phrases. Google did it out of a partnership with Samaritans back in 2010. Maybe they have yet to push Bing to include the feature?

But it’s hard to argue that Bing wins this round of the search battle by giving the user exactly what they want – even if what they want is information on how to end their own life.

Now search for “porn.” Nothing, right?

In some instances, Google does this to protect copyright. For instance, searchers of the word “torrent” will find a dearth of autocomplete suggestions. Just recently, Google added The Pirate Bay to its list of blacklisted search terms for autocomplete. In some cases, words like “amateur,” “porn,” “boobs,” and other related terms are blacklisted to…really I don’t know. To protect Google users from *gasp* pornography, I guess.

One of the words that failed to produce any autocomplete suggestions was “bisexual.” Notice the past tense here. That’s because a bisexual advocacy group is claiming to have won the battle and gotten “bisexual” off the blacklist.

“Since late 2009, Google has had “bisexual” on a list of banned words; such words were de-prioritized by the Google search algorithm, leading to a drop in search rankings for all bisexual organizations and community resources. Since its search engine would not auto-suggest or auto-complete any term with the word “bisexual”, Google made it harder for any user to find bisexual content, whether that be on coming out as bisexual or finding local support groups across the United States and elsewhere.” explains BiNet, a longtime bisexual advocacy organization.

They are claiming victory in an effort to change that. The group’s head, Faith Cheltenham, had this to say on her personal blog:

Google search results WILL vary by user, one user has already reported seeing “bisexual” when typing in “bi” while other users don’t even see “bisexual quotes” when typing in “bisexual q”. The block was lifted on August 21st as far as some VERY DEDICATED volunteers can tell. It was on that date that “bisexual q” started producing “bisexual quotes”. Just a few weeks later, I get these results when typing in “bisexual q”.

For me, typing “bise” produces a suggestion for “bisexual quotes.” I’m not seeing any other suggestions for “bisexual” on its own or anything else, however – but Google’s instant suggestions are different for everybody.

What’s particularly odd about this is that for years, Google didn’t block autocomplete results for words like “heterosexual,” “homosexual,” “asexual,” and even “trisexual.”

“We thank Google for making the right call here and for acting as a responsive corporate citizen committed to dignity and equality,” said Kate Kendell, Director of the National Center for Lesbian Rights.

It’s important to note that Google has always been one of the most vocal companies in their support of LGBT rights.

[via Slate]