ara pehlivanian

In January I wrote about the apparent holes in Gmail’s spam filter, and since then the problem hasn’t gotten any better, in fact manually reporting spam has become a new sport for me–and not one that I’m particularly fond of.

Today, right in the middle of a chat with Mathieu, something bubbled up from my subconsciousness and sparked a sneaking suspicion in me that Gmail may be deliberately using me as a human spam filter. In fact I’m almost positive they are. There’s no other way to explain that many “lottery winnings” notices leaking through the filter when even I can write my own filter in Gmail–using their own filter tool–to catch and throw most of them in the trash. Just looking for the keywords “lottery, win,” and “congratulations” or any combination of the three will block 90% of them alone. So they must be deliberately letting spam into my inbox in order to get a human to confirm it. What else can explain it?

Add to that the talk given by Luis von Ahn, at the Googleplex, called Human Computation where he talks about a game he created that gets unsuspecting players to tag images on the web while having fun, and it becomes clear that the same thing is being done here–except it isn’t fun. Whose to say that the spam filter guys at Google weren’t sitting in on the talk and aren’t trying some variation of the technique in Gmail? I mean c’mon, Akismet–which I love–behaves like a proper learning spam filter, Gmail behaves like a drunken bouncer at a night club flirting with all the girls waiting in line while everyone else sneaks in behind him through the front door.