Julian Sanchez header image 2

photos by Lara Shipley

Why Yahoo’s “Occupy Wall Street” Block Actually Matters

September 20th, 2011 · 2 Comments

Yahoo found itself at the focus of some brief fuss today, after their e-mail service’s spam and malware filters started blocking many emails that contained the phrase “Occupy Wall Street” or linking OccupyWallStreet.org in connection with an ongoing protest that had attracted some 3,000–5,000 people over the weekend, with a few hundred die-hards remaining on Tuesday as the protest entered its fourth day.

The conspiratorially minded hinted that this might be overt political censorship, which was presumably good for clicks, but also completely ludicrous if you thought about it for ten seconds or so. Try to imagine the conversation:

Suit: Hey, Steve, listen, I know it’s sorta winding down, but I’m starting to worry this little Wall Street protest could spell the end of our nefarious system of capitalist exploitation. So I want you to start giving people error messages when they try to send legitimate e-mails about the protest. Sure, it’ll piss off our user base, and it’ll probably be terrible PR, but it’s a small price to pay for crushing the Revolution. We’ll fix the block within a day or so and blame it on a “technical error.”


Tech: But, sir, listen… most active Internet users these days have a second e-mail account, not to mention Twitter, Facebook, IM, Google+, blogs… won’t blocking people’s e-mails in such an obvious way just annoy our users for no reason, and draw more attention to the protest? I mean, the blocking in Egypt was a lot more comprehensive and sophisticated, and they still ended up having to shut down the whole Internet—do you really think one provider’s crude filter is going to make that much difference in a protest that’s already been so widely announced? Also, uh, I’m no lawyer, but political censorship doesn’t sound like a “routine business purpose.” Even if our lackeys in the corporate press can be persuaded to look the other way, aren’t we opening ourselves to a federal felony prosecution under the Electronic Communications Privacy Act if anyone figures out we did this on purpose? What exactly is the benefit to Yahoo! of inviting a potential shitstorm just to block a tiny percentage of emails about one more demonstration?


Suit: Hell’s bells, Steve, where’s your ruling class solidarity? Your career, my career, even the company itself—all well worth the risk if our dear friends at Goldman Sachs have one less hippie to wade through on the way to work!


Here’s an ever so slightly more plausible scenario. There’s suddenly a huge uptick in e-mail traffic, maybe a lot of basically identical messages CCed to large numbers of users by enthusiastic activists, with a link to a very recently registered domain. Lots of recipients who got a copy from vaguely-recalled acquaintances blasting their address books flag it as spam, and (whoops) the Yahoo filter algorithm decides the URL is a spam marker. Also, given that Anonymous has been one of the more vocal champions of this particular protest, it’s not exactly beyond the realm of possibility that some overeager supporters might have tried to do some kind of automated mass notification. Let’s tentatively say that this is at least as plausible as the whole pointless-and-risky-conspiracy-that-makes-no-sense theory.

So this is a “move along, nothing to see here” kind of story? Ah, not at all! The second scenario is actually still pretty disturbing: It means a bunch of legitimate political speech got blocked on the basis of some very superficial similarities to mass emails that the algorithm might well have reasonably pegged as spam. That’s still a fail on Yahoo’s part, and a sign they really ought to take another look at their detection algorithm.

Still more worrisome, though, is the possibility that if they don’t fix it, this incident makes it clear that it’s possible to exploit Yahoo’s algorithmic overkill to block third-party messages containing certain phrases or (especially) links. In this case, it was probably an inadvertent, distributed own-goal scored by people whose sincere attempt to promote a cause backfired. It could just as easily be exploited deliberately by someone sending pseudo-spam seeming to promote a cause they actually wanted to harm. In a way, this would be a modern variant of a classic political dirty trick: A campaign starts putting out offensive flyers or annoying, repetitive dinnertime robocalls purporting to come from their opposition—and the bewildered opposition suffers the backlash. Imagine what an unscrupulous political operative could do in the 48 hours before election day with a false-flag spam campaign? Not only do they pepper voters with hard-to-attribute misinformation, but as a lagniappe, they manage to block the opposing campaign and its supporters from sending genuine get-out-the-vote messages!

I guarantee you that somewhere out there, the Karl Rove of the Social Media Douchebag set has already thought of this—sits steepling his or her fingers and emitting a guttural “eeeeexcellent” at the prospect. That gives Yahoo about 13 months to fix the problem before it resurfaces in a way that might actually make a difference.

Tags: Tech and Tech Policy


       

 

2 responses so far ↓

  • 1 K.Chen // Sep 21, 2011 at 10:52 am

    “Karl Rove of the Social Media Douchebag set” is probably my second most favorite thing you’ve written.

    Couldn’t the problem be ameliorated in part by the multitude of e-mail service providers using presumably different propriety filters? And, separately, I recall the Postmaster General talking about the idea of the USPS getting into secure/authenticated e-mail delivery. Maybe this will spur that sort of development

  • 2 Spam Sentries Shoot First And Ask Questions Later | Component Parts // Sep 21, 2011 at 11:46 pm

    […] Sanchez thought this didn’t sound plausible, reasoning that it was more likely the URL somehow got flagged […]