Troubling Solution to a Real Problem
Jonathan Zittrain
This essay previously appeared in U.S. News & World Report. It also appeared in the Internet Monitor project’s second annual report, Internet Monitor 2014: Reflections on the Digital World. The report, published by the Berkman Center for Internet & Society, is a collection of roughly three dozen short contributions that highlight and discuss some of the most compelling events and trends in the digitally networked environment over the past year.
2014 saw an extraordinary move by the European Court of Justice to compel search engines like Google and Bing to allow individuals to demand the removal of certain search results returned by their names. Whether or not you support such a right in theory, the ECJ’s chosen implementation raises several big problems.
First, should the search engine decide not to act upon the request, the person making the request can appeal — but there is no parallel mechanism for anyone else, such as the affected website, to appeal when the search engine grants a request. Moreover, when that happens, a removal appears to be permanent, even though changing circumstances might make an initial removal no longer warranted. For example, the requester might become a public figure, making previously irrelevant information vitally relevant. Or the contents of a webpage such as a Wikipedia entry might be changed to remove or amend the information in controversy. Or the requester could die, affecting his or her rights to the sorts of reputation shaping the ECJ contemplates. Affected sites should have a formal path to appeal search engine takedowns, and the grantings of takedown requests should be for a limited time, with an opportunity for the requester to re-lodge a request before expiration of a previous successful takedown.
Second, especially because the takedowns themselves are currently unreported to all but an affected website — and even that notification is, in some quarters, being challenged by authorities — there will be no way to analyze how the new right is being invoked and granted. There should be a means for academics willing to respect subjects’ privacy to study takedowns, discern patterns, and report in aggregate how the right is evolving and what sorts of requests are being granted versus denied.
We are likely to see more and more automated requests, with some people using specially developed intermediary reputation services that will perform rapid searches on the users’ names, automatically categorizing results as negative, neutral, or positive, and then acting as the users’ agents to file automatic takedown requests. This could result in a large volume of requests untouched by human hands, including those of the requester — likely not an outcome contemplated by the European court.
Finally, there is a puzzling triumphalism among some who view the ECJ’s decision as a victory for European sovereignty. It’s true that the decision compels Google, Microsoft, and others to consider EU law in rendering search results, and perhaps under European pressure, around the world rather than simply through European-branded versions of their products. But the core holding is a mandate for Google and others to hand-tweak results, breaching a longstanding editorial-style barrier that reputable search engines have taken pride in maintaining. For example, Google for many years has offered an explanation to those curious, and at times outraged, to find that a search on the word “Jew” often points to an anti-Semitic site as its first hit. That explanation — http://www.google.com/explanation.html — indicates an unwillingness by Google to hand-adjust search results to suit particular tastes, including its own. Rather, Google undertakes broad changes to its search algorithm in the name of increasing relevance or combatting spam and other search engine optimization. To ask Google and others to adopt the habit of adjusting search results for very specific outcomes risks more broadly breaching the editorial barrier that has served as one of the public’s best defenses against propaganda by centralized services like Google and Bing. It would be an irony if the very decision seen as re-asserting regulatory sovereignty over intermediaries instead unleashed far more self-conscious, but still secret, information shaping by them.
With luck, 2015 will see arrangements between the major search engines and academic institutions, especially European ones, to facilitate study of these phenomena, so that a right created to serve the public interest can be evaluated against it. Without these arrangements, I can only join others in having a hunch in how it’s playing out — and mine is that the ECJ has come up with a bad solution to a very real problem.
