With some fanfare, Twitter announced yesterday that they were implementing a mechanism for IP-based censorship of tweets on a country by country basis -- triggered by demands made by individual countries themselves.

It's been accurately noted that Twitter (like most other global Internet firms, including Google and Microsoft) has long globally removed certain materials as a perceived legally required response to individual countries' "orders" in these respects. Localized takedowns (such as the removal of certain Nazi-oriented search results in France) are also not uncommon.

And obviously such companies must obey these orders if they are to continue doing business in those nations -- even when the orders are utterly contrary to basic tenets of human rights and free speech.

In their analysis of this situation, EFF briefly mentions -- but seems to downplay -- the "slippery slope" aspect of country by country censorship capabilities. EFF also asserts that "transparency" requirements -- e.g., Twitter publicly noting when takedowns occur, represents a significant mitigating factor.

But Twitter's new acceptance of localized censorship perhaps provides an opportune moment to take another look at this entire situation, with an eye toward determining whether such approaches do more harm than good, both globally and locally.

Unfortunately, there is virtually no evidence to contradict, and vast evidence to support, the notion that the more "localized" and "frictionless" a censorship system, the more governments will expand their use of such systems over time.

In the Internet context, the problems triggered by providing localized censorship capabilities are easily visible on both sides of the fence.

For the country doing the censoring, localized censorship capabilities are a dream come true. They can push through all manner of takedown demands, knowing that pushback from the Internet services involved will be minimized by the fact that such services no longer have to risk the global wrath of users in other countries affected by global takedowns. So especially for those countries who already imprison their citizens for merely writing a blog post or sending a tweet that is not in keeping with the party line, the local censorship approach seems a page directly from their oppressive playbooks. And they can feel pretty confident that those firms won't feel the need to cease providing all services to such countries, since they can censor in those nations without affecting the remainder of their global user bases.

Local censorship can indeed sometimes be easier on the Web service firms. They are no longer faced with the binary choice of accepting global censorship demands, vs. pushing back on such demands (as they would likely be more inclined to do when global takedowns are the norm), and perhaps having to threaten to no longer do business in or provide services to those countries.

Essentially, localized censorship systems provided by global Internet firms can enable a mechanism for evading key foundational ethical questions. If faced with a total cutoff of such important services, rather than an easy way to consummate their censorship demands, would many nations actually choose the former, total cutoff course? How would their populations react? What would be the political and other ramifications?

But these become purely academic questions when country by country blocking and takedowns are available, because the systems then provide a comparatively frictionless path for both governments and Web service providers. The losers, by contrast, are the ordinary people in those censored countries.

It is certainly true that some knowledgeable users in such countries will likely find ways around blocks and associated nationally-based takedowns, through proxies or other means. But most persons in those nations probably will not know how to do so, and would be fearful of the risks associated with being discovered using such circumvention technologies.

Nor is there significant evidence suggesting that transparency "listing" requirements reduce censorship at all. Most purveyors of censorship have little concern regarding who knows that they are engaging in censorship, so long as the censorship is taking place.

I would never assert that this is a simple situation, or that these are easy decisions for the firms involved. There are many complex and frequently conflicting aspects, forces, and considerations in play, both in purely business and other spheres.

Still, my gut feeling is that moves toward localized censorship enabling systems ultimately fall into an ethically suspect arena. While they likely reduce the need for global takedowns, they also seem destined to perpetuate and over time increase the use of censorship against national populations by their governments, sometimes in intensely oppressive manners.

There is a disquieting sense of trying to make peace by pushing censorship behind borders where everyone else doesn't have to deal with it, while the targets of censorship are left to wither under an ever escalating barrage of national censorship demands.

It's difficult to see how the scourge of censorship can be realistically fought in such an environment, where for all practical purposes the result of these arrangements is censorship being effectively given a stamp of approval, so long as any given item isn't censored on a global scope.

This may indeed be a workable mechanism for avoiding the need to fight intense anti-censorship wars, or threatening countries with major service cutoffs.

On the other hand, if we actually care about persons under the boots of censorship loving regimes, perhaps we should be giving a bit more consideration to the notion that some concepts -- freedom of speech among them -- truly are worth fighting for.

--Lauren--