In the last two decades there has been much talk about how artificial intelligence is modelled after the brain — what with terms such as “neural networks” and “deep learning” commonly used to describe some features of how AI works.

But some scientists with expertise in both neuroscience and AI suggest that AI may also teach us how the brain works. 

Up to now, most profitable work on the brain has been based on simple systems, simple species and therefore simple brains, and in more complex animals, studying more manageable parts of the brain such as the visual system from retina to the sensory cortex, the motor system from the motor cortex to motor neurons in the spinal cord, the auditory system from the inner ear to the temporal lobe, the olfactory system, somatosensory system including touch-pressure, thermal and pain sensations from the periphery to the sensory cortex, memory, position sense, transmission in nerve fibers and synaptic transmission and so on — all manageable bite-sized pieces of the mammalian nervous system. 

Perhaps the biggest bugaboo of all has been consciousness: one of those all-embracing but enigmatic words that means very different things to physicians, neurophysiologists, psychologists and even physicists such as Penrose and Schrodinger.

Francis Crick, who won a Nobel Prize for his work in DNA, spent the later decades of his life wrestling with the neural nature of consciousness — to little avail. 

Most people equate consciousness with awareness of self, others and the environment in a meaningful manner, which adds more words but little clarity.

Some scientists, such as Oliver Sachs, thought questions about the consciousness weren’t worth the effort, that it was a pit for all except philosophers and that the question would quietly disappear as we better understood how the brain operated in general. Maybe so.

As I write this essay on AI and the brain, I’m aware of writing with my right hand, steading the notebook with my left hand and a sentence-by-sentence flow of ideas, but little else, except the garden when I look up from time to time — but little else.

That may be what I’m aware of, but while I’m pondering and writing, my brain receives countless messages from sensory receptors, to update it about what’s going on in the world outside and within the body, all of which I’m blissfully unaware, thank goodness.

I’m also aware that if I change my thought or what I’m aware of and my focus shifts, it’s almost impossible — for me at least — to hold two or more separate thoughts simultaneously.

It’s somewhat akin to visualizing numbers in my mind: when the next number pops into view in my mind, the previous one disappears.

You may be better at this game than I am, but it makes a point, when our attention and awareness changes, what last held our attention tends to fade away, at least momentarily. 

Our brain is capable of nimbly shifting what it attends to and what held our attention moments ago drops below our attention threshold. This pattern of shifting attention goes on all the time.

Some temporary attention holders may not be so temporary: the brain continues to work on them out of our awareness and what was temporary a few minutes, hours or even days before, may pop back into our awareness — sometimes with solutions to whatever we were fussing about.

That’s were AI comes in.

Like the human brain, AI acquires large databases (sets for AI), which cover a broad range of subjects sequestered in different sets.

The issue is: how does AI choose from among different sets of data to better focus and solve the problem or question at hand?

Some have likened this process to small staff meetings in a large company with a lot of different data sets offering specialized and sometimes competing information for attention. 

That’s where intelligence comes in: managing and making best use of large data sets and picking and choosing which set is most appropriate to a given task, without losing touch with other data sets.

That’s also what AI is learning to do better. It’s also what the brain effortlessly does — most of the time. 

For example, much of the primary information from the sensory, motor and emotional realms come together in the association cortex in the parietal, temporal and frontal lobes.

The latter provide integrated and potentially actional information that may prompt us to become aware of whatever situation is going on, while tamping down competing information below the threshold of our awareness. 

That’s probably how the brain operates and, in like fashion, AI operates.

Is the latter conscious and/or aware? Probably — if it walks like a duck, it is a duck.

The difference is that one, the brain, is carbon-based and the other, AI, is silicon-based and, incidentally, one down from carbon in the periodic table.  

So much for consciousness and awareness — although I know I haven’t heard the end of this debate from colleagues.

Dr. William Brown is a professor of neurology at McMaster University and co-founder of the InfoHealth series at the Niagara-on-the-Lake Public Library.