Once again, I begin by apologetically mentioning the long hiatus between posts here. I suppose it’s only fitting that I break the silence by musing about whether the silence has really been broken. If a blogger blogs on a deserted blog and nobody reads it, has he made a virtual sound? This musing about the impact of random ranting has been triggered by a couple of recent incidents, one widely publicized and one more obscure.
Anyone familiar with recent events in national news may have guessed one of the topics. George Sodini recently opened fire at a fitness center in Pennsylvania, killing several people and wounding more. As police and news media investigated, they discovered his blog. As far back as November, he had been posting bitter messages about his unhappy life, and lack of success with women. Several entries indicated vague plans for his grand finale, including messages before and after a previous attempt in January when he “chickened out”.
One article I read said that police wanted to know if anybody had been “monitoring” his blog, and if so, why authorities had not been notified. I think the answer to that is pretty obvious. The net is full of millions of blogs read only by their owners. Sodini had no friends. Nobody cared what he had to say.
I suppose at this point I could muse about a never-married, middle-aged computer geek with a lousy sex life and a blog that nobody reads. But that’s not really where I want to go with this.
Instead, I’m going to contrast the unsurprising fact that Sodini’s rants and threats remained unread with a slightly more surprising incident. Consider this
recent discussion on a bulletin board which apparently is visible to everyone, although it requires registration before posting. It obviously attracts more attention than Sodini’s soliloquies, even though some of its attention may be computer-assisted.
At 7:34 AM on 7/28, in a discussion on what people are cooking for dinner, one member goes off on a tangent about her broken refrigerator and problems with Sears repair service. At 1:48 PM the next day, 7/29, a message appears in the discussion from someone posting with the screen name “searscares”. This screen name was just registered that day, obviously for the specific purpose of responding to this discussion. It belongs to a senior case manager for the “Sears Cares Team”, who is disappointed with the customer’s troubles and wants to rectify things.
This is an amazing example of data mining. Nobody notices Sodini’s not-so-subtle threats to shoot women because he can’t get laid, until the damage has been done. Yet somebody on an obscure bulletin board casually mentions that she can’t get her refrigerator fixed, and within 30 hours, the Sears Cares Team is on the case. When I told my boss that threats to kill people get far less attention than gripes about Sears refrigerators, he wondered what would happen if somebody threatened to kill people and store their bodies in a Sears refrigerator.
I had previously read that many companies were beginning to proactively search some popular “gripe sites” for complaints, to try to nip PR disasters in the bud. But I had no idea that they had been able to extend their ears this far. As one person mentioned later in the discussion, it might even be a little creepy. I’m just curious about the filters and AI that must be at work here, to mechanically crawl through billions of words every day and successfully extract a manageable number of items for review by humans to decide if further action is necessary. I’d love to see some numbers on how much data gets processed daily by their computers, how many items get flagged for review, and how many of those turn out to be legitimate issues. If I say “My Sears refrigerator sucks and the repair crews are morons”, maybe they will see this and contact me and answer my questions. Actually, I have no Sears appliances, not because I have anything against Sears, but just because I haven’t bought anything from them. I do think they deserve kudos for such a proactive attempt to make their customers happy.
So if Sears can find customers with broken refrigerators, could, or should, law enforcement agencies apply similar techniques to find nutcases with broken hearts and loaded weapons? Although I’m a strong proponent of civil liberties and privacy rights, I can’t see how anybody could reasonably object to the scrutiny of information that’s widely available to anyone who takes the time to read it. I don’t want them reading people’s email or tapping phones without proper cause and legal warrants, but anything that’s bared to the whole world is fair game. I’m constantly amused by people who post embarrassing or incriminating information on publicly-available websites, and then scream about privacy violations when it’s read by law enforcement or employers. In some cases where people have been fired over such issues, I think the employers have over-reacted to a trivial issue and the employee has a legitimate bitch about that. But it’s preposterous to stretch the issue and claim that their privacy was violated by reading something that was made available to anybody with a browser.
So if it’s acceptable for law enforcement agencies to read anything that anybody else can read , is it feasible to troll the web for signs of danger? Well, it’s already being done to some extent. Anybody who posts an online threat against the President has a good chance of getting a visit from the Secret Service. If I made such a threat here, would they find it? I’m not going to try. I don’t want to try to explain to people with Uzis that I was just testing to see if they care as much as Sears (although I might not object to a body-frisk from that hot brunette that was on GW’s security team in Lexington a few years ago, as long as it doesn’t involve a Taser).
But it probably really isn’t feasible to try to expand this kind of surveillance to more random threats. And even if it was, I’m not sure it’s desirable. I think there is a definite possibility for abuse. Again, I acknowledge that it’s acceptable to read anything that is posted publicly. But reading and reacting are different issues. Somebody has to make a judgment call as to whether what they read constitutes a legitimate threat that justifies action. And thanks to the ubiquity of cell-phone video cameras, the eagerness of some police officers to escalate issues is disturbing. It is unfortunate that Sodini’s blog wasn’t read by even one person that could have said “Um … maybe this guy needs to be watched.” But how wacko does somebody have to seem before you judge that they’re a threat and not just a harmless kook? I can’t answer that, and I’m scared to think how it might be answered by some.