The Chinese government is notorious for censoring content on the web; to do so, it has created a cyber-security mechanism known colloquially as the Great Firewall. But in truth, many world governments are looking to control the information shown online in one way or another. An extraordinary set of data published today by Google (NSDQ: GOOG) gives insight into this process.
In a major update to its Transparency Report, Google has displayed 18 months of data that shows how many “content removal requests” it has received from government authorities in a variety of countries. In the United States, for instance, courts and cops made 182 separate requests to remove content from a Google service. These requests cover all Google services: YouTube videos, Gmail, Blogger blog posts, and of course search results.
Other companies that have content online—especially user-generated content—are surely receiving their fair share of these types of requests as well. But none, so far, have gone as far as Google in making public information about these requests. This isn’t the only example of Google sharing data about who’s trying to control information online; it also is a major contributor to Chilling Effects, an online repository of copyright and trademark takedown requests.
Overall, the report is a fascinating snapshot of how the world’s governments are reacting to the internet.
The part of the report that deals with “content removal requests” has the most interesting details:
» Brazil is the leader when it comes to demanding that Google remove online content. That’s because it’s the one country where Google has had great success with its Orkut social-networking site, so Google simply has a lot more of people’s personal data posted online in Brazil than in other countries.
» A few agencies with many requests. Often, Google performs repeated takedowns in response to requests from a particular government agency. For example, in Brazil, electoral courts fired up during the fall elections season ordered takedowns of content on Orkut. In South Korea, an information-security agency routinely requests removals of search results that contain RRNs, a personal ID number assigned by the South Korean government. Google complied with all 139 of the requests from South Korea that it received from July to December 2010. In Germany, a government youth protection agency routinely asks for Google to remove URLs from search results when they contain “Nazi memorabilia, or extreme violence or pornography.” Google complied with 97% of the 118 requests sent from Germany. In Thailand, the Ministry of Information sent one request for removal of 43 pieces of content that mocked the king, an act which is illegal in Thailand; Google disabled those items, but only for Thai users.
» Some big cases. Some individual requests lead to big swaths of information being taken down. In the U.K., the Office of Fair Trading requested the removal of fraudulent Google ads placed by scammers. Google took down 93,360 “items” due to that request. In the U.S., six court orders stemming from one ongoing defamation case led to the takedown of 1,110 items from Google Groups.
» Bad apples? While Google complied with most content removal requests, some countries kept striking out. Libya sent 68 requests for content to be removed, but Google complied only 31% of the time. In India, Google complied with only 22% of the 67 requests it got. Some of those requests were from “different law enforcement agencies to remove a blog and YouTube videos that were critical of Chief Ministers and senior officials of different states,” the company states. “We did not comply with these requests.”
It’s worth noting that a “content removal request” can be a lot of different things. It could be a classic kind of government content-regulation, like France or Germany’s laws prohibiting hate speech, or Thailand’s ban on talking smack about their king. Or it might be a court order sent to Google after a plaintiff wins a defamation case.
» More countries are demanding takedowns. In 2010, Google started receiving removal requests from many countries that had previously not made such requests. Those newcomers to the internet-control party include Cyprus, Greece, Hong Kong, Kazakhstan, Libya, Portugal, Taiwan, Turkey, Panama, Mexico, Argentina, Croatia, Denmark, Hungary, among others. In other countries the number of requests increased significantly. See Google’s country-by-country notes on trends that it’s noticing.
» The China question. The Chinese government is arguably the world’s most notorious internet censor—and it isn’t even on Google’s list. Why? Because the censorship demands themselves are state secrets. Google cannot reveal any of the censorship orders it received until June 2010. Beginning in June 2010, Google essentially shuttered its Chinese search engine at google.cn, instead directing users to its Hong Kong site at google.com.hk, where it can provide uncensored search results.
So now Google is free to disclose censorship demands sent by China—but the requests have stopped coming. In the last six months of 2010, there were no such requests.
» What’s not in the data. Google takes down search results and other content connected to child pornography on an ongoing basis; sometimes that’s as the result of a government request, and other times it just does it on its own. Those takedowns aren’t in this data. Nor are, for the most part, copyright takedown requests, since they (nearly always) come at the behest of private parties and not governments.
Some interesting points from the section of the report dealing with government demands for user data:
» Governments want more user data from Google. Google received close to 9,000 requests for user data from the U.S. government last year, far more than it received from any other country. Google complied with 94% of those requests, according to data from the second half of the year. (Google only started monitoring how many requests were complied with in July 2010.) Unlike content removal requests, most of the demands for user data come from developed nations. The data requests are generally connected to criminal investigations.
But the company’s non-compliance with 6% of requests from U.S. law enforcement—and a much higher proportion in other countries—suggests that it’s seeing a lot of overreach in the data that cops around the world want to grab about suspects.
» A few countries aren’t getting the info they want from Google. For a few countries, Google rebuffs most or even all of their demands for user data. The company complied with only 12% of the 272 user data requests from Polish authorities; none of the 68 requests originating in Hungary; and none of the 45 requests from Turkey.
But for most countries, Google complies with a high proportion of data requests. In the U.K., it’s 72%; for Israel, 76%; for India, 79%; for Brazil, 76%; Australia, 81%;
After the U.S., the countries with the most requests for user data are Brazil, with 4,235 requests in 2010; and India, with 3,129 requests in the same period. The large number for India may just be an effect of its large population. The high number for Brazil is probably due in part to the fact that it’s the one country where Google has had good success in the social networking area, with its Orkut network. That means the company is holding a trove of Facebook-type personal data that it doesn’t have in other nations.