Google and censorship — setting the record straight

Oh no, someone is wrong on the internet again! In this particular case, there’s been a slew of articles lately implying google is complying with government demands for censorship when that isn’t the case, so I need to set the record straight.

Cheney’s House:

This story seems to repeat once a year: This year Sharon Weinberger in Wired’s Danger Room blog got the ball rolling when she asked why imagery of the vice president’s house is blurred on Google Earth — or more precisely, asking why Google Earth is blurring the vice president’s house.

The blog eventually got a response from Google — that it sometimes uses imagery that is already blurred by the source, and that other mapping providers like Yahoo and Microsoft have similar imagery — but the headline of the post, “Why is Google Earth Hiding Dick Cheney’s House?” remains unchanged, implying agency where there is none.

I’ve asked people at Google familiar with the matter about this: If the specific imagery of Observatory Hill provided by the USGS and used by Google has been blurred due to of out-dated and out-of-place cold-war concerns, why not just get an unblurred version in the public domain via other channels, like Ask.com has done?

houseask.jpg

At the very least, it would prevent bloggers from bothering Google spokespeople every couple of months asking about Dick Cheney’s house. The answer is that Google is looking for replacement data, but getting the licensing rights to publish it to the web can be tricky and/or expensive. That’s a far cry from a conspiracy to censor.

51 things:

Also in July, ITSecurity came out with a list of “51 Things You Aren’t Allowed to See on Google Maps” that got Dugg and saw widespread re-reporting; and here too, agency is implied where there is none:

Whether it’s due to government restrictions, personal-privacy lawsuits or mistakes, Google Maps has slapped a “Prohibited” sign on the following 51 places.

But of the 51 items posted, in only one case did Google actively roll back imagery for security reasons at a government’s behest — in Basra, Iraq in January 2007. (Street View imagery removed because it was mistakenly taken from private property is not interesting from a censorship perspective.)

In the cases of blurred bases in the Netherlands and blurred energy sites or research labs in the US: that’s because the aerial imagery used was censored by local authorities before being released to the public. Google could, if it wanted to, use satellite imagery from other sources that show these areas unblurred, but chose not to, either for cost reasons or because it figured that the quality of the higher resolution aerial imagery offset the downside of having some specific sites blurred. Again, no agency on the part of Google.

So why couldn’t Google offer multiple imagery datasets — say a base dataset with 15m imagery plus DigitalGlobe’s content, and a dataset for higher resolution aerial imagery that might be censored in places? I’m sure that it could if it had an unlimited budget and unlimited development resources — and in that case, there are many more other items on my wish list: a historical database of datasets, for example. It’s just that Google doesn’t have unlimited resources, and it isn’t obliged to spend those resources it has on actively circumventing pre-existing censorship.

In other cases, ITSecurity’s list is just plain wrong: I don’t know why this meme continues to persist, but Google has not censored any imagery in India (see item #24) — I had this confirmed to me again recently. Nor has Yona in Guam been censored (item #27). It’s easy to make a list like this if accuracy in reporting is not important. It certainly doesn’t help that there are a lot of inaccurate articles out there on the web being returned in search results and subsequently referenced. Still, it doesn’t reflect well on the credibility of ITSecurity, which is ostensibly a serious news site.

Raw Story:

The Raw Story’s John Byrne, in an article from Aug 26 entitled “Google Earth increasingly compliant with censorship requests: US intelligence report” uncritically re-reports ITSecurity’s “list” as factual but also rewrites a report by the US government’s Open Source Center from July 30, 2008. Surprise surprise, the report is rife with factual errors:

After the Basra incident, Google Earth seemingly became more open to dealing directly with foreign governments to assuage their security concerns. It agreed to blot out British bases in Iraq and other sensitive UK installations such as the eavesdropping base at Cheltenham and the Trident nuclear submarine pens in Faslane, Scotland.

No it did not. Nor is there any evidence of imagery being blurred at the Chinese government’s request, as the report insinuates. (Satellite imagery is not available to those inside the great Chinese firewall, but that isn’t the rest of the world’s problem.)

The Raw Story’s Byrne re-reports as fact further errors from ITSecurity’s “list”:

Among the areas Google blurs out in China includes, not surprisingly, Tibet/Xinjiang Province. Other areas of Asia that have been clouded include northern areas of Pakistan — it’s unknown why or who might have requested the omission.

That’s just a ridiculous statement. Does anyone even bother to fact check any more? Tibet, Xinjiang and Pakistan’s Northern Areas are riddled with high resolution imagery squares from DigitalGlobe. Where is the censorship? Are these people so clueless that they assume that areas without high resolution imagery are censored?

Is it a big deal that these stories are out there on the web? The main problem is that people read them uncritically, assume they are factually correct, and sometimes even re-report this fiction further as fact, as informationWeek and Strategy Page did. Eventually, it becomes the received wisdom that governments can get their way with Google, when the reality is very different. I abhor censorship of my mapping data as much as anyone (if not more) but I think it is very important to sift actual cases from urban myths, and in general the web is failing dismally at this task.

3 thoughts on “Google and censorship — setting the record straight”

  1. The answer lies somewhere between the nature of these contracts, which I doubt anyone will be able to access, and the nature of the industry, which is fractured and very un-standardized at best.

    Many data providers will simply not cut deals that aren’t all-inclusive of their data, or, in some cases, which compete partially or wholly with Google’s other data providers. In other words, Google (or any other consumer-level service provider) is only as good as the data they can get. And, yep, even Google can’t get everything it wants. (This is also why they are now starting to roll out their own tools, their own map coverage, etc. among other obvious business reasons.)

    As for the latter, blurry homes are consistent with the patchwork bird’s-eye tilesets Google (or anyone else) serves, reflecting an industry that isn’t very standardized in terms of display (say, versus data formats).

    As LBS become more and more prevalent, we are going to see a lot of shake out in both data providers and how the data is provided. No doubt.

  2. I want to correct you on the Washington, DC imagery.

    The 2005 aerial photography of Washington, DC released by the USGS was censored more by Digital Globe than by the USGS. Although this was highlighted before, Google has not updated the downtown area with the newer USGS imagery. In order to see this imagery, you have go through the USGS and download the imagery of downtown Washington, DC. In this case, its more about passive censorship through obstructing the ability of users to see the newest available public domain imagery and Google is complicit.

Comments are closed.