The trouble with AI

Your brain works in a wonderful way. The things your eyes see, the senses of touch, the hearing – all these are data input traveling along nerves as electric signals, on or off, strong or weak. The brain then receives this data and organizes it along patterns it recognizes from learning since birth and stores it as engrams — analogue interpretation of data.

Let’s take an office digital scan of a photograph. As the light and light-receptor passes over the image, individual points of data are “seen” and “recorded”—passed as digital impulses into a data file to then be reconstructed—dot by dot—on your computer screen or piece of paper (ink dots applied one at a time). At no time does the scanner, computer, screen or printer recognize the whole image. But when you look at that image … your brain remembers the whole, all at one time.

The digital world is seeing a painting as brush strokes, each one true, not seeing the whole. An analogue world is seeing the whole and then, if asked, seeing/recognizing what it is made up of, stroke by stroke.

Now let’s jump to artificial intelligence, or AI. The problem with AI is that it is digital, seeing only data points as truth, storing not the image but the data points and the rules/code to reassemble those data points to reproduce (not recognizing) the whole. 

But this is always an incomplete understanding — if one can call it an understanding at all. Yes, a programmer can tell AI that the painting is beautiful, but the AI cannot mentally “see” the painting, just the data points precisely because it is a digital machine, not an analogue brain storing wholeness.


Because of this flaw, AI can ruin civilization. AI will make decisions based on data, not understanding, not replicating humanity but abiding by values and rules set by civilization, yes, but unable to see the moral, analogue, understanding of those values and rules. The more we empower AI to run the commerce and infrastructure of civilization, the more a digital interpretation of right and wrong will be applied. 

A woman approaching a police check point, visualized and analyzed by AI, could conclude that all the data points of her attire, pace, facial expression, hair color, shoelaces — all these and more do not individually pose a threat. 

A human might see that the whole of her human posture/visual impression seen as an analogue comparison of “normal” might mean she is wired with a bomb. 

Digital AI cannot protect humanity, such AI cannot act responsibly for humanity. Such AI — controlling everything from supply-side commerce to medical needs — could destroy civilization and with it all humanity.

However, AI that is analogue — possibly with quantum computers — should  be able to recognize/mimic humanity and therefore could still take over civilization but not destroy it since it’s very existence would be dependent on the analogue tenets and engrams of civilization with and by which it was created.


Peter Riva, a former resident of Amenia Union, now lives in New Mexico.