Intelligent Machines?

Not To Worry. . .

Worrywarts have come to love the Sunday New York Times because there’s often fresh stuff to worry about and yesterday’s issue was no exception, offering the worry that machines may come to be smarter than people.

The article by John Markoff nicely details how far machine intelligence has come and where some fear it may be going, including concern that “further advances could create profound social disruptions and even have dangerous consequences.”

Although anyone who has wrestled with one of the phone machines used by service companies to avoid giving service may debate the point, the skill and autonomy of machines is growing.

Even detractors admit that many machines are as bright as freshman Members of Congress, and cost less besides.

Still, there are many signs that machines have yet to develop intelligence equal to humans and as a public service, some top examples are listed here, a roll call of recent human achievement that even the best and brightest machines could never duplicate:

n War In Iraq. No machine could ever have made the decision to go war with Iraq at a cost of lives in the thousands and treasure in the trillions to eliminate weapons of mass destruction that weren’t there. And no machine could have taken more than six years to swap one corrupt, family-and-faction-centered government for another corrupt family-and-faction centered government. Proud results of human, mostly American, intelligence on display for all to see.

n Worldwide Recession. No machine could have so steadfastly looked the other way while New Age Rumplestiltskins spun gold not just from straw, but from manure in the form of securities backed by securities. When the manure finally began to stink more than the economy could bear, only human intelligence could have bailed out the spinners and returned them to their task.

n American Health Care. No machine could have devised a system where rising costs deliver declining results, except for the 40-50 million who live – and often, die – outside of the system’s reach. And no machine could persuade many of those who are in the system that change would be an assault on their precious right to pay more for less.

Beyond these great recent human achievements are countless little things that only a human could do. Have you ever seen a machine smoking? Or drunk? Or texting while driving?

Or how about a machine that hates on the basis of color or language or gender? Or one that consumes more energy than it needs? Not a single widget has mastered even the fundamentals of these uniquely human and simple tasks. Humans master most of them effortlessly while yet children.

No machine on its own has yet to murder, lie, cheat or steal. Only ages of evolution could have developed those skills so easily embraced by the human intellect that yet elude machines.

Bottom line: In spite of concerns raised in the Times article, machines learning to think like humans isn’t a legitimate concern.

No machine would be so dumb.

###