Algorithms versus depression

June 21, 2010 by · Leave a Comment
Filed under: Uncategorized 

An Israeli researcher, Yair Neuman, has developed an algorithm that analyses blogs and identifies depressed bloggers based on their writing. When clinical psychologists were asked to evaluate the same blogs, their views and the software’s output matched in 80% of the cases. Neuman sees this development has having practical applications: for instance, by enabling mental health workers to identify individuals who might need psychotherapeutic help.

Neuman himself seems well aware of the privacy issues here. He had permission from all of the bloggers whose writings were included in the study to use their blog entries; also, in his description of a usage scenario of his software he says, “Through this software it will be possible to contact a blogger and request a general examination of the contents of his blog. If the blogger agrees, he will know whether he needs to seek professional counseling for any possible distress”.

There are a number of interesting issues here. First off, one might say that this has no place in a blog about privacy as people’s blogs are inherently public. If someone is blogging, they want the world to know what they have to say. However, they may not have intended their writing to be analysed for the sake of making some kind of psychological report.

Second, one might say that a blogger who wants privacy can blog under a pseudonym. Third, one might say that an algorithm that can tell me that someone who wrote “oh my god my life is not worth living i’m so depressed” is depressed is not that clever. We don’t know which texts were analysed, so we’ll have to withhold judgment on that.

If there is a privacy issue here – and I’m pretty sure that there is – it is related to the technologies I wrote about in a previous post, namely, technologies that claim to infer our psychological state of mind from remotely taken physiological readings. If this technology could create some kind of psychological profile of a blogger, even, or rather especially, when he is not writing about how he feels (but rather about sport, say, or technology and privacy), then it would be a very powerful tool. I am not saying that this is what Neuman’s algorithm can do right now; instead I’m using it to think about possible developments, one of which might be the creation of a psychological profile of a blogger that stretches over years (the internet never forgets, remember).