Google, the Benevolent Behemoth?

June 18, 2012

google censored
Google's new transparency report documents an alarming rise in censorship by governments, from the US to China.

If you are not familiar with Google’s transparency reporting, you should be.

By monitoring access to Google services and publishing that data in real time, Google’s transparency tool “visualizes disruption in the free flow of information, whether it’s a government blocking information or a cable being cut,” which has great potential to augment early warning efforts for mass repression.

At any time, you can see requests for url removal from search results for copyright claims, and see who those purported owners are. As we know from discussion on this blog around PIPA and SOPA, Google’s efforts to combat infringement of intellectual property rights—at least narrowly defined—are in keeping with human rights law, and important for staving off really bad policies.

Google states in the report that 'we hope these steps toward greater transparency will help inform ongoing discussions about the appropriate scope and authority of content regulation online.' I read this as an unambiguous cry for help.
Requests from governments to either effectively remove content (by making it unsearchable to those using Google at their computers or smartphones), and requests to provide user information to governments is published by Google every 6 months, however.

Yesterday, Google released a snapshot of data from the last 6 months of 2011. The topline: “Google reports ‘alarming’ rise in censorship by governments”.  In an accompanying blog post, Google told of a request by the Spanish government to remove 270 search results linked to blogs and news related to public figures as an example of the types of increasing requests it receives.

The reports provide country-level data on government requests for removal of content, with differentiation between “court orders” and “other requests” such as from police or executive agents. And while not comprehensive—Google does not provide data from countries where there are fewer than 10 content removals—both the data itself, and the decision to publish the information are public goods.

The Good and Bad of the Public goods

There are two very important pieces of information in these transparency reports on government requests. The first is the frequency of request for user data or (non-copyright related) content removal. The second is Google’s compliance frequency.

The first could tell us quite a bit about the strategies of repression used by certain governments, especially when matched against other indicators of such repression.

But the second—Google’s compliance—tells us little, save for the smallest insight into Google’s compliance policies.  Despite the wealth of information in the data, we are stuck with an inferential problem.

Take, for example, the case of Brazil. In the 6 month period ending Dec 31 2011, Google fully or partially complied with 90% of government data requests, compared to 64% of data requests from the UK government. According to Google, the volume of requests in Brazil is a function of their use of the orkut social networking site.

But what accounts for the difference in Google’s response rates? One could infer that the UK government makes more flimsy requests than does the Brazilian government; and that may or may not be true. But another explanation is that the criteria used by Google are different. And they are. Different jurisdictions have different laws relating to defamation, hate speech, and pornography, among others.

That is a human rights compliance problem. Whether the requests are made by court order or via executive agents, some countries simply do not have the necessary legal safeguards or rule of law to guarantee freedom of expression. In some places, political or religious speech is highly regulated in clear violation of human rights law. What does Google do in these cases?

Google states in the report that “we hope these steps toward greater transparency will help inform ongoing discussions about the appropriate scope and authority of content regulation online.” I read this as an unambiguous cry for help.

As I’ve offered with regard to Twitter on this blog, Google as a multinational corporation in the business of information exchange will be faced with a steady barrage of government requests to censor information, and provide user data.

Google is being confronted with removal or user requests that comport with a country’s domestic law, but is wholly incompatible with human rights law. Should Google grant that request, human rights watchdogs must call it out, as we would with any corporation that facilitates human rights abuses.

To that end, I’d like to offer that Google’s compliance with 70% of Thai requests to remove monarch-insulting content is troubling (and thanks to Google for the data). Given our reliance on Google’s self-reporting, a classic moral hazard is upon us.

To be absolutely clear, we should be very pleased to see Google’s transparency reports, and the broader initiative, as we should to see that Google is fighting back against censorship and government monitoring in China.

But at least as it relates to basic guarantees of speech and expression, the dramatic expansion of the internet and networked communities is forcing us to identify a common denominator. I wager that Google doesn’t want to set it, and despite the promising moves by the corporation so far, I wager we don’t want them to either.

As an aside, the US government made–by far–the greatest number of user information requests (for over 12,200 user accounts). Google complied with 93% of requests. Again, a sincere thanks to google for providing the data.

If you’re reading this from the US, take action to pressure passage of the Global Online Freedom Act.