Back in 1996 IBM’s Deep Blue made the news when it beat world chess champion Gary Kasparov. While very impressive that a machine could learn to play chess, the ultimate point wasn’t to make a great chess playing computer. Like more recent game playing programs, such as AlphaGo, the real objective is to build smarter & smarter machines that can learn, adapt, and think for themselves. The real goal is the advancement of artificial intelligence (AI).
Early computers were programmed by humans by hand to do specific jobs - pretty straightforward. A calculator was programmed by humans on how to calculate numbers and that’s what it did. But as we have added layers of complexity with greater demands, computers have become so complex that it is no longer possible for humans to program all aspects of some programs. Instead, humans are building parts of programs that go off and build other parts of programs themselves and find the most optimized logic through trial & error. As machines are programming parts of themselves, a human trying to explain exactly how a program is doing what it is doing is actually becoming un-explainable.
So as machines are getting smarter they are being given more complicated jobs - jobs that were strictly the domain of humans. Computers are learning to drive vehicles, read medical test results and make treatment recommendations, write news articles, and of course manufacture goods. One of the benefits a machine has over a human is that, for some of these tasks, a machine is (or will be) better at the job than a human. Another benefit is of course that machines don’t take sick days, they don’t need to sleep, and can do the work of several workers at once. The implication is that AI is coming for your job.
For the field of business intelligence it should be obvious that with time, computers will become much better at data analysis than humans. It’s the kind of field that seems perfect for AI - virtual mountains of data at the disposal of AI programs to learn from, analyze, and find the best most reliable algorithms to make recommendations. We already use BI software for its ability to sort through large data sets at lightning speed. The next step is for machines to actually analyze the data and not just retrieve information when called upon.