Anima Books
books by holistic veterinarian Dr Christine King
Miscellany
Remember who you are
The faux supremacy of "AI"
Recently, I read an article lauding the use of commercial “AI” software (SignalPET) in veterinary radiology:
The premise of the study was that software such as this may be a good substitute for board-certified veterinary radiologists, making expert radiological assessment available to more vets in-house (i.e., in the veterinarian’s clinic) and at a lower cost:
Sounds good, right?
The trouble is that the machine missed about 1 in 3 abnormal radiographs (x-rays), calling them normal when the consensus among the veterinary radiologists in the study — all specialists in their field — was that the radiographs in question were abnormal.
For us science nerds, the “AI” software’s sensitivity (true-positive rate) was only 68.8%, making its false-negative rate 31.2%. Its false-positive rate was much better, 5.6%. With a specificity (true-negative rate) of 94.4%, it mistakenly called a normal radiograph abnormal only about 1 in 18 times.
In short, the software was inferior to the veterinary specialists its designers were aiming to replicate — or should that be replace?
“AI” was inferior to the specialists its designers
were aiming to replicate — or should that be replace?
It’s worth noting that the first two authors plus the fourth are economists, not veterinarians, and all three are affiliated with the Max Planck Institute for Research on Collective Goods. (Is that the same as “the common good”? How Orwellian that sounds…) Only two of the five authors have any veterinary affiliations (vet schools in the UK and Germany).
Of course, the authors concluded that this use of “AI” in veterinary medicine is a collective good:
But it’s a con to call this software “intelligence” and to consider its output radiological interpretation (the phrase used in the paper’s title) or diagnosis. It is simply performing programmed pattern recognition — and rather poorly at that.
It is incapable of interpreting the meaning of any radiographic pattern in the particular patient’s clinical evaluation and subsequent care.
It is incapable of interpreting the meaning of any pattern
in a patient’s clinical evaluation and subsequent care.
In fact, it has no idea that the image is of a living, breathing animal to whom a person has attached value of some sort, whether physical (e.g., service or working dog), emotional (e.g., family pet), or merely financial (e.g., racing, breeding).
It knows very little. It understands even less. Most importantly, it does not know us, in all our warm and wriggling complexity, so it cannot take the place of us.
It knows very little. It understands even less.
It does not know us, so it cannot take the place of us.
Who are we in relation to this novel tech?
So, who are we in relation to this novel tech?
A few months ago, I gave a workshop titled Medical Intuition in Veterinary Practice — blending science and intuition for better patient care. [I’m in the process of expanding it into a book, so stay tuned.]
In the first part, I went through the various functions or processes specific to the left and right cerebral hemispheres — i.e., ‘left-brain’ and ‘right-brain’ characteristics. I’ve summarized them in the following table.
What struck me the other day is that “AI” appears to be designed to replicate and enhance, even supercharge, the functions or processes of the left-brain exclusively. It appears to have no capacity at all to replicate those of the right-brain.
It’s as if the software developers — and that’s all “AI” really is: a bunch of computer programs, designed by human software engineers (although increasingly by “AI” itself) — have completely disregarded the other half of their brains, and thus their capacity and their potential as human beings.
It’s as if the developers completely disregarded
the other half of their brains, and thus their capacity
and their potential as human beings.
Incidentally, an earlier version of the table, based on the work of neuropsychiatrist Mona Lisa Schulz MD, PhD, also listed the left-hemisphere processes or perspective as masculine and those of the right hemisphere as feminine. It’s not lost on me how hypermasculine the “tech bros” try to present themselves and encourage other men to emulate.
This is what sets us apart from “AI,” which I put in quotes for a reason: while it is assuredly artificial, it is not in any way intelligent as I think of intelligence (an innate characteristic of all living systems that allows them not only to survive, but to thrive).
This is who we are in relation to this novel tech.
This tech appears to completely ignore an entire hemisphere of our potential as individuals, groups, societies, and as a species.
Missing from the dystopian, “AI”-controlled future that is fast bearing down on us, whether we like it or not, is the supremacy of our ability not only to think and reason and act for ourselves, but also to:
* imagine
* be inspired
* be awed
* be curious
* be intuitive
These abilities are common to us all, even though Western civilization has for at least the past 400 years prioritized the functions of the left-brain, relegating those of the right-brain to the province of the artist, and all the value judgements that go along with it.
It’s no wonder that “AI” so closely replicates the functions of the left-brain and pirates those of the right-brain to generate works of art and music that are entirely derivative of those already created by human beings.
So, my message to all of us, myself included, is this:
Remember who you are, and revel in the fact that “AI” can never even come close to replicating or replacing the magnificent you.
Remember who you are, and revel in the fact that “AI” can never
even come close to replicating or replacing the magnificent you.