The Ideas piece, by Emily Kumler, The bias in the machine (July 9), states, Typically, a programmer instructs a machine with a series of commands, and the computer follows along.

This statement captures in broad stokes the larger contours of the here and now of computing and artificial intelligence though far from entirely so, of course. Reality isnt that lock-step the computer [slavishly] following along with a series of commands. To that point, the essay further assumes a straight-line development of AI, such that whats expected longer term is more sophisticated programming leading to still-genuflecting computer obedience.

Advertisement

The future of AI, however, will likely be very different than that. Rather, AI will depend decreasingly on human intervention for its thinking and increasingly on its self-programming, as machines learn more and more heuristically. That is, the trajectory of AI systems will be to independently acquire, curate, adapt, and apply knowledge in order to inform and shape and reshape its own behaviors and eventually to do so, if human egos can relinquish some of AIs executive functions, far more competently than erstwhile human programmers.

Keith Tidman

Bethesda, Md.

Go here to read the rest:

The 'bias' of artificial intelligence - The Boston Globe