I’ve been becoming increasingly bemused by the masterful example of technology marketing that is IBM Watson.
It’s a great (and rare) example of a tech company building a strong brand around a very esoteric set of concepts. Just about everything coming out of Big Blue seems to have the Watson brand slapped across it at the moment. Ask anyone what Watson actually is and you’ll get a vague “Jeopardy-playing, AI ML Big Data robot lawyer thingy”.
The anthropomorphic naming is a masterstroke. Giving something human attributes through giving it a human name is far more powerful than any of Watson’s no doubt impressive algorithms. But as someone who has a vague understanding of things technical I have the deep sense that we are being duped.
It strikes me that the AI/machine learning/big data triumvirate have three main purposes today:
– the processing of mass data to tell us things about today- language translation or behind research, for example. To a great extent I’d call this “simulated intelligence”. It’s not actually “thinking” to any great extent, just using vast amounts of processing power to crunch data.
– Turing Test-type fake human stuff that you can have a chat with. This is probably where the general perception of AI lives in the greater population (if such a thing exists at all). From what I’ve heard, this is a bit of a sideline these days as simulated technology approaches have taken over as AI increasingly doesn’t worry about meaning as it gets great results from inference from data. Let’s call this “Artificially Intelligent agents” or something similar. But I’d put Siri, Cortana and Alexa into the “simulated” category.
– predicting the future stuff, using data to foresee things. The IT industry has been promising this for years – witness stalwarts of computing from earlier decades Delphi, Oracle and Sage. Let’s call this branch “bullshit”. Unfortunately this is the one that people really want, because most of us crave to remove uncertainty and doubt from our lives.
Oh, to predict the future. IBM Watson is named after Thomas Watson, the man credited with making IBM the brand it is. He made predictions. In 1942 he predicted that the world market for computers would get to about 5. That’s 5 as in one more than 4.
And here is my beef with the “use the data” to predict the future crew. The only meaningful predictions that you can make based on the data of the past is that the future will probably be a little bit like today, no matter how much intelligence you throw at it. And most of the time you’ll be right. Except when you aren’t, at which point you’ll look like a bit of an arse. A bit like IBM’s original Watson.
So remember that next time you encounter the Watson marketing machine rears its head. Watson 2.0, from the people who brought you the prediction of 5.
mmitII 