Google and Microsoft should be open about their de-linking processes in Europe

Like it or not — I’m somewhere in the middle — Google(s goog) and other search engines operating in Europe have to take down links to information about people if those people ask them to do so, provided there’s no public-interest or other good reason for keeping the links up. Now we’re in a stage of flux, where everyone’s trying to figure out how to achieve this in the best way possible. In Germany, for example, there’s talk of setting up new cyber courts to judge de-linking requests.
The problem for now, however, is that the search engines are already plowing ahead with the takedowns and no-one is sure how exactly they’re doing so.
We first saw how things can go horribly wrong when Google removed links to items that were clearly in the public interest. It did a u-turn on some of those, but it still never said why it had taken them down in the first place. Now Microsoft(s msft) has also started taking requests — and like Google, it refuses to give a clear picture of how it judges requests.

The full picture

To be fair, Google legal chief David Drummond did give a partial view into the factors his company takes into account, in his Guardian article on the subject a week ago:

“These include whether the information relates to a politician, celebrity or other public figure; if the material comes from a reputable news source, and how recent it is; whether it involves political speech; questions of professional conduct that might be relevant to consumers; the involvement of criminal convictions that are not yet “spent”; and if the information is being published by a government.”

So we have the ingredients but not the recipe — we know what the criteria are, roughly, but we don’t know which criteria Google weights above others. That’s a big problem because, as Drummond himself said, “these will always be difficult and debatable judgments.” You can’t debate something you can’t see.
The problem is serious because it’s hard to push for a link to be reinstated once it has been taken down. In the case of those erroneously deleted BBC and Guardian links a couple weeks back, the publications went very public once Google notified them of the takedowns, but that won’t always be the case, and the links will usually just vanish with little attention.
By way of contrast, if Google or Microsoft refuse to act on a takedown request, that can be appealed with the national data protection authority. In other words, there’s legal recourse in such situations, but not when the decision goes the other way.

“No-one safeguards the public interest”

As Lilian Edwards, Strathclyde University’s professor of internet law, told me by email, referring to the search engines:

“They are indeed not a court nor a public body, so FOI [freedom of information] and judicial review both don’t apply – this is always the problem with the internet. To be honest, Google are better than most; they have said right-to-be-forgotten takedowns will go into their transparency report and they’ve put together a respectable academic panel to oversee ethics. But my argument throughout has been that no-one safeguards the public interest in this. Data protection authorities may respond to complaints of failure to take down, but it is not in my opinion their remit to examine when takedown should NOT have occurred.”

As I’ve stated before, Google and other search engines should not have to act like courts, particularly when they’re facing tens of thousands of requests. They obviously shouldn’t automate the takedown process DMCA-style, but nor should they have to decide borderline cases themselves – if there’s even a grain of doubt, they should reject the application and leave it to the applicant to go complain to the local data protection authority, if they feel strongly enough about it.
But if the search engines are taking it upon themselves to play judge – which the ruling at the heart of this did not say they must do – then the public deserves to know with some precision how these choices are being made. I fully appreciate that these methods are a work in progress, but that should not stop Google and Microsoft from providing running commentary on how they are evolving. As Edwards said to me, there isn’t a legal blocker in the way here.
It’s not just me asking for this – European data protection officials, who are due to meet with Microsoft and hopefully Google too next week, also want the search engines’ test for removing data to be more transparent. I reckon this could be accomplished with a series of blog posts. We don’t need to know the details of every case (information overload won’t help anyone, and it would defeat the purpose of the exercise anyway) but there’s no reason we can’t get further insight into what the takedown teams are thinking and how they are operating.
It may look messy and it may keep changing, but it’s a reality that the public deserves to see, just as surely as the public deserves to know how their real courts operate.