It takes a lot to get Google to deindex a page, and thus hide it from searchers (at least from U.S.-based searchers). Unlike with YouTube, where Google exercises considerable editorial discretion, Google Search is generally aimed at indexing the Web, good and bad. Until recently, there have been only a few categories of content that Google would deindex based on someone's request (setting aside Google deindexing things itself based on a perception that someone is gaming its search algorithms, or that some site contains malware, and focusing just on Google search within the U.S.):
- Legal obligation (mostly copyright): U.S. copyright law threatens to hold search engines liable for linking to infringing material. Google therefore acts on DMCA takedown requests from people who claim to own a copyrighted work that a site allegedly infringes, though there's also a counternotification procedure that the site operator can use to get the material brought back up pending a copyright lawsuit.
- Confidential personally identifying information (such as social security numbers) or revenge porn: Google has no legal obligation to deindex pages containing such material, but it voluntarily does so, again based on just the objector's request. In the U.S., this is not a general right to be forgotten—it's limited to content like a "security or government ID number, bank account or credit card number, an image of your handwritten signature, [or] a nude or sexually explicit image or video of you that's been distributed without your consent."
- Court orders addressed to third parties (chiefly in libel cases): Here too Google generally has no legal obligation to deindex material—if you sue me for libel and get an injunction finding that my page libeled you and that I must remove it, that generally doesn't bind Google (though the Hassell v. Bird case now pending before the California Supreme Court might affect that). But if you send the order to Google, Google will consider voluntarily deindexing my pages, because of the court finding that the pages are libelous. (There are risks here, chiefly because many people have submitted forged and fraudulent orders; but that's a separate matter.)
It has just emerged, though, that Google has decided to deindex based on a fourth category:
- Administrative agency findings that sites illegally distribute material that risks physical harm to consumers: Right now, this category appears to include just warning letters from the U.S. Food and Drug Administrations, generally sent to off-shore online pharmacies that illegally sell prescription drugs to the U.S. A Google representative told me that this is supposed to be a narrow policy, limited to fact findings by "administrative agencies that are charged with protecting consumers' physical safety from harm by products or services that they consume," where there is reason to think that there was "some process through which site operator had notice and opportunity to be heard" in the investigation. The policy does not extend to agency findings of financial harm or reputational harm; as item 3 indicates, Google may act on a court order that certain material was, for instance, libelous, but factfinding by an administrative agency wouldn't suffice. But where there's an administrative agency finding that material can be physically harmful to consumers, a court order is no longer required.
Google told me that it's possible this may extend to other agencies, but so far they "have not received any removal requests from anyone but the FDA." They are currently contemplating only orders from federal agencies (and similar national agencies in other countries), because the effect of the deindexing would be nationwide.
So far, the deindexing requests that I've seen (see, e.g., here) have focused on sites that really are just online pharmaceutical stores. If a site appears to have more material (e.g., general information about pharmaceutical efficacy, political advocacy, and the like), then Google would try to deindex only the material that is illegal and that threatens physical safety.
Google and the FDA both state that this is an entirely voluntary policy on Google's part, and not motivated by any threat of enforcement against Google (and I doubt that there would be any current law that could yield such a threat of enforcement against Google just for providing links).
* * *
So those are, as best I can tell, the facts. What should we think about them? I don't think there are any viable legal objections to Google's new policy. The government apparently hasn't pressured Google into implementing it, so there are no First Amendment or Due Process Clause problems. And there's no law requiring Google to link to anything. (Indeed, such a law constraining Google's selection decisions might itself be unconstitutional, as Don Falk and I argued in a 2012 paper commissioned by Google; but in any case, no such law exists.) The questions here are ethical, prudential, and business questions.
So let me just flag some of these questions, without offering an answer.
Should Google be praised for protecting the public from the dangers of unprescribed—and, potentially, adulterated or otherwise harmful—pharmaceuticals?
Should Google be faulted for restricting what its users can see, just for their own supposed good and not to protect any concretely harmed third parties?
Should we worry that, just as category 3 (court-order-based takedowns, including for reputational or economic harms) has led to category 4 (administrative-finding-based takedowns, at least to prevent physical harms), category 4 will eventually slip into much broader takedown practices, based just on an agency's say-so?
Should we be glad to see such slippage, at least to prevent what appears to be criminal conduct?