SOMEWHERE ON THE INTERNET - “Will you write an article about what might happen ‘when artificial intelligence (AI) outsmarts humans?’” they asked. “Of course,” I thought, “that’s going to be fun: It can be a flight of fancy into a future that has so many possible strands that the possibilities are endless.” Then I engaged my brain.

When conducting any form of inquiry, there are always ‘the five Ws’ - the who, where, what, why and when - and all good writing starts with that query. In this case, however, there are two more ‘Ws’ - and they are both ‘will.’ To be more specific, will AI outsmart humans? And can AI exercise free will?

This article will examine not AI - for that has already been done in previous articles - but rather the concept of ‘outsmarting.’ To do this, one must also decide what one means by ‘smart.’ One often says that people who live on their wits are smart. By that, one means that they learn what they are taught and learn from both what they experience and what they are exposed to.

Can machines do that? Yes, that is what is called ‘machine learning.’ Does it, however, qualify as ‘intelligence’? No, it does not. It is neither artificially intelligent nor intelligent in any other way. This is because machines do not actually learn in the sense that sentient beings do, nor do they have genetic memory. They only know what they have been taught or what they have ‘learned’ in passing.

Still, ‘smart’ is often used in other ways. One, almost obviously, is the parent of the ‘living on one’s wits’ way. It refers to the using of information - however that is defined or derived - to produce or predict results. It is also the using of the absence of information because that absence is also information in and of itself.

Riding light

There are advocates for what one could call algorithmic analysis that clai