The world is becoming, for better or worse, increasingly algorithmic.
Recently a friend, who was at work at the time, came back to his desk to see a message similar to this one on his iPhone: ‘We think there may be a problem with the connection between your Jawbone and our servers, because it reports that you have been in the same place for the last four hours’.
He had been stationary for the least four hours – he had been on a conference call, an occupational hazard for the corporate financier. But this reflects the logic of the machine at work; if the location stays the same for that length of time, and it’s not night time so they are likely not sleeping, a problem in the relaying of location data may very well be the most likely cause. To the individual who knows the context, however, this conclusion is humorous precisely because it belays how little the algorithm knows about the real world.
The autumn issue of 2600 magazine included an intriguing article by an ex-employee of Comcast who worked in technical support. The gist is that the employees are given a list of questions and answers, and that is it. If the question is not on the list, no answer is given. This approach is so well known about, at least in technically literate circles, that it has become a kind of cliché. A commenter on Slashdot stated: “The Comcast phone slaves won’t have a page on their script to fix his problem. Might I suggest pulling the power plug from the router and rebooting the PC, though.” Comcast’s approach begs the question: why have staff for this function at all, given that this function can be fulfilled just as adequately by a document available on the internet?
Another example which will be familiar to people who have bought things on Amazon occurs when you look at an item and purchase it or put it in your basket as a reminder to purchase. You then continue around the internet and adverts pop up for the thing you have been looking at. Of course, by this point these adverts are superfluous because you have already decided to purchase it.
Those who are familiar with technology are often the most sceptical of new technological solutions, because they are familiar with the uncanny ability of each solution to beget new and often unexpected problems. Not all algorithms are smart.