Artificial intelligence will not be like human intelligence. Already, the way computers think is very different from the way people think. Computerized “brains” are constrained by logic, whereas human minds are rational only very selectively. Machines have different capabilities from humans — different areas of expertise — and they are designed to approach problems from alien angles.
We may reach a point when artificial intelligence looks like human intelligence. It will be programmed to mimic our mannerisms and to present human-seeming ideas. But when machines become sentient, they will surprise us. They surprise us now! Imagine how strange and foreign their creative abilities will be.
Current computers perform labor that resembles creativity, but we say their output is not truly novel because they were programmed by humans. I wonder if we need to interrogate the concept of “creativity” — after all, humans come with biological presets as broad as instincts and as specific as nucleotide arrangements. Are we anything but squishy supercomputers? Answer: no, not so much.
Written after listening to the latest episode of the Exponent podcast, “OpenAI and Strategy Credits”. Also posted as a response to “Superintelligence Now!” by Steven Johnson.
Comments are closed.