The geopolitics of Google's autocomplete

Google's autocomplete algorithm doesn't just enable users to save precious seconds of typing by predictavely filling in the rest of the search. It's also, apparently, the subject of contentious legal cases the world over. The latest example: On Wednesday, a German federal court ruled that libelous autocompletes are a violation of privacy.

As the BBC reports, the case was brought by a businessman (fittingly, he remains unnamed) who was frustrated by the fact that autocompleted searches of him with "scientology" and "fraud." This week's ruling -- which overturned two previous decisions in favor of Google -- called on the search giant to make changes to its autocomplete function when made aware of an "unlawful violation."

And this is far from an isolated case. The BBC goes on to report:

The ruling could also have a bearing on another case involving auto-complete. Bettina Wulff, wife of former German president Christian Wulff, sued Google because auto-complete suggested words linking her to escort services. Mrs Wulff denies ever working as a prostitute and has fought several legal cases over the accusation. The case against Google is due to be heard soon in a Hamburg court.

The technology blog Techdirt, which snarkily claims to have a "suing-algorithms-for-fun-and-profit! dept" brought us another story last year of an Australian surgeon named Guy Hingston who sued Google for defaming him by implying that he's not doing so well financially. The search:

But as TechDirt pointed out, Hingston may be shooting himself in the foot. His case, in attracting media attention, has made it all the more likely that "bankrupt" will appear next to his name in a search.

In 2012, ZDNet wrote about a Hong Kong tycoon who sued Google for similar reasons. As ZDNet noted, "Whether Yeung's name is input into Google Search in English or Chinese, a drop-down option for the search term plus 'triad' [the name for China's organized crime organizations] appears -- a connotation which is unlikely to make the tycoon happy."

And individuals aren't the only parties bringing autocomplete-related lawsuits. In 2012, an anti-discrimination group in France, SOS Racisme, sued Google for discriminatory autocompletes -- in this particular instance, linking "Jew" or "Jewish" with searches for people who aren't Jewish like Rupert Murdoch. Go figure.

With so many loose associations on Google, does it really make sense to hold the company accountable for each one? After all, you could argue that everything from women to countless countries to former British Prime Minister Gordon Brown have been defamed by autocomplete. Google, for its part, claims little responsibility. Their defense: the algorithm works by filling in blanks based on the frequency of our searches. In other words, we're all kind of slandering each other.

Screenshot [h/t Telegraph Online]


The New Yorker takes a page from WikiLeaks

WikiLeaks may have launched as a venue for publishing sensitive leaked information, but it eventually began working more closely with established media outlets to showcase its scoops.

Today, with the unveiling of its new Strongbox feature, the New Yorker is taking the next step, cutting out the middleman entirely and going straight to the whistleblower: Strongbox is a system similar to WikiLeaks's dropbox where sources can provide information to the magazine's reporters and editors through a system that can't record IP addresses, browsers, computers, or operating systems.

Strongbox uses the Tor network (designed by some of our FP Global Thinkers!) to ensure I.P. address anonymity, and provides users with randomly generated code names they can use to sign in (you can read more about how it works here).

The New Yorker actually isn't the first news organization to adopt the WikiLeaks model -- the Wall Street Journal tried something similar in 2011, though some noted at the time that the site had technical issues that could compromise anonymity (as AllThingD points out, the Journal hasn't said a lot about how much use it's gotten out of the site). The code for Strongbox, written by the famed late programmer/activist Aaron Swartz, is open source, which means we might well see other news organizations set up their own dropboxes in the near future (Swartz was working with the investigations editor at Wired to put the project together, for example).

In one sense, Strongbox isn't quite breaking new ground: sources have been leaking anonymously to news organizations ever since there was wrongdoing and people around to write about it. But as this week has shown, the protection of that anonymity has become more difficult -- even when news organizations try their hardest to maintain the privacy of sources. The secure dropbox was part of what initially made WikiLeaks innovative. Could it transform news reporting as well?