Machine intelligence: Scary? Necessary.

We owe a lot of rich drama and philosophy to the concept of machine intelligence. From the pure horror (or, uh, campy comedy?) robot overlords inspire, through Asimov’s deep probing of the nature of humanity, to the idea of the Singularity – rephrased as “Transcendence” in the upcoming movie of that name.

But what does an intelligent machine necessarily “transcend” (other than “the threshold of intelligence” – which is tautological)? Mechanisms – profoundly complex mechanisms, to be sure, that elude our current understanding – underlie human intelligence. Indeed, the human brain is routinely likened to a wonderful machine. Why should a machine that succeeds in performing a feat we normally associate with intelligence – such as consuming data from a wide variety of sources, “remembering” (storing) that data in a form it can readily access and correlate, and using what it has “learned” to arrive at a likely answer to a question of fact  – not be characterized as intelligent?

I submit that many “machines” (human-designed systems) display intelligence, today. To be sure, it is a far less flexible and remarkable intelligence than you, dear human, have displayed in arriving at this blog post and (I hope) making sense of it, in the context of your own background, interests, and motivations. It is also far less mysterious. I believe it is merely the fact that we do not (yet) understand how we are so intelligent that makes most of us believe – nay, insist – that we are far more intelligent than anything we design shall ever be.

Whatever you believe, as an information technology professional I trust you will agree that the “smarter” the system we can create for our client, the better. Let’s further agree that the output of a “smart” system is:

  • clear,
  • accurate,
  • timely,
  • consistent,
  • trusted, and
  • pertinent; and therefore
  • more valuable than the input

That sets the bar nice and low for “machine intelligence” – or does it?

In future posts, I’ll share some of the best practices I’ve discovered for creating and maintaining intelligent “information machines”. Meanwhile, if this post inspires or provokes you, I’m sure you’ll let me know about it!