2023-11-03 21:42:18
by MASSIMO SIDERI
With the newspaper the first of three titles in the series «The human side of technology» created with the Humane Technology Lab of the Catholic University. Thinking Machines and Interaction: Kate Crawford Explores the Dark Side of AI
In the second half of the nineteenth century the case of the intelligence of the horse broke out in the media of the time. It went down in history as the “Cleaver Hans” dilemma, from the name of the horse, Hans, and from cleaver, intelligent, and we can put it in these terms: can the expectations of the observer or the person asking the questions influence the answers? In short: yes. Through the intelligence of animals we have understood a flaw in human intelligence.
Even today, as University of Berkeley professor Kate Crawford recalls in her book Neither intelligent nor artificial, the dilemma is used to evaluate the responses and prejudices, this time positive, that we ourselves cultivate towards algorithms. Here’s how it works: Hans amazed the world with his ability to count and perform simple mathematical operations, such as additions; if he was asked, for example, how much is 2 plus 3, he would respond with 5 taps of his hoof.
The case fueled public imagination and expectations so much that the German Ministry of Education set up a commission to understand whether its owner, Wilhelm von Osten, was a fraud. But nothing led to confirm this hypothesis. There was no scam.
Yet there was a trap, of which everyone was the victim, starting from the horse and its owner. Hans used what today we would call “emotional intelligence” rather than the ability to calculate: at the fifth blow of his hoof, or the seventh, he was able to notice the satisfaction of the person asking him the question and stopped. It was a psychologist who discovered it: Carl Stumpf (he understood the subtle mechanism of subtraction: when someone who didn’t know the answer asked the questions, Hans’ intelligence collapsed). An ante litteram example of the importance of interdisciplinarity.
Historia magistra innovationae, history also teaches us about innovation, how to study it. The anecdote about Hans the horse is in fact useful for asking today the question that the father of artificial intelligence, Alan Turing, asked himself in 1950: can machines think?
Our very expectation that they can do so leads us to believe it. It is not just a clue to the mechanism of “faith” we place in technology. There is something deeper and more unconscious. A bit like when our mind captures anthropomorphic features in a cloud, or when it seems to us that our cat or dog behaves just like a human being. Sigmund Freud called it the uncanny, thus giving a name to that ambiguous sensation of attraction and repulsion that we feel when we look at something inanimate that however reminds us of our life: a mechanical doll in his time, an android robot today.
In our case it is a phenomenon which, with a neologism, we could define as «AIcebo», a sort of placebo effect of artificial intelligence (AI, or in English AI): our anxiety to see signs of intelligence in machines pushes us to find even signs of consciousness. As with experiments with placebo substances and molecules, the effect arises within our world, at a crossroads between psychology, the chemistry of life and biology.
Fiction is not new and was also seeded in the nineteenth century: it was a contemporary of Charles Darwin, Samuel Butler, who published a book hiding behind the pseudonym Cellarius entitled Darwin among the machines. In the story, halfway between futuristic non-fiction and novel, Butler anticipated a world in which machines would acquire awareness and take control of the world by evolving Darwinianly.
Here is the seed of «AIcebo». It is as if our mind had been “hacked”, partly by our need to explore and discover new geographies, partly by the advancing marketing army of technological companies that would like to colonize our very idea of the future.
In philosophy it is called the paradox of the ship of Theseus: according to the myth, to commemorate the escape of Theseus from Crete with the young people freed from the prison of the Minotaur’s labyrinth, every year the Athenians went in procession with the ship used by Theseus. The paradox, told by Plutarch, is this: after a hundred years, when the pieces of the ship have been removed one by one to be changed due to obsolescence, is the ship still the same?
The paradox was taken up and complicated, centuries later, by Thomas Hobbes: if someone took the individual pieces dismantled from the ship and made a second ship, is this the original? In other words, is it the form or the material that guarantees identity? We could apply the same thing to the human mind: it is as if, in the interaction with technology from the wheel to AI, we had changed the individual neural pieces while leaving our mind in the same place. In this sense our brain has been “hacked”.
At Carnegie Mellon University, one of the many tests was organized to try to quantify the fragmentation of the ability to concentrate when we live with technology: 136 boys and girls were divided into two groups. The offline tribe, i.e. with the phone turned off, achieved an average of 20% better results in various tests. The others lost. The cost of distraction. The experiment is a further example of how we automatically but unconsciously put ourselves at a disadvantage in the (incorrect because passive) use of technology.
There is also an exponential effect: when we lose concentration we don’t just lose the amount of time dedicated to something else, for example checking a notification. Our mind has to refocus, go back a bit, think back to what it read before. The process of wasting time is not the only sum of the fragments of time lost. Like when we pick up a book left on the bedside table after a while. Where were we? Indeed: what was the story about? What were the characters’ names? Who was this Starbuck? And this Edmond Dantès? You have to flip through a few pages, back and forth. It’s the exponential cost of distraction. The human side of technology, ultimately, is not about asking whether machines are intelligent. But how can we use them to remain intelligent sapiens.
The series in three releases and the meeting in Milan
Kate Crawford in the book Neither Intelligent Nor Artificial. From the first pages, The Dark Side of AI uses the metaphor of the atlas to describe her work. The volume, which has a preface by Antonella Marchetti, is on newsstands for €9.90 plus the price of the newspaper, the first of three issues in the series «The human side of technology» published by «Corriere» and edited by the Humane Technology Lab of Catholic University directed by Giuseppe Riva. «Maps in their best form – writes Crawford, visiting distinguished professor at Berkeley but also co-founder of the Ai Now Institute at New York University – offer us a compendium of open paths, ways of shared knowledge, which can be mixed and combined to create new interconnections.” The book is also a map of the journeys that the author has made to carry out the investigation: in the Amazon factories as well as in the “farms” where the human beings themselves are paid a few cents to tag images on the web to allow AI to discern and therefore “appear intelligent”. On the other hand, the critical nature of Crawford’s work (who is also a senior researcher at Microsoft, thus showing a freedom of thought that is not obvious) is evident right from the subtitle and that dark side. Like all maps, her ambition is to help human beings find their way back. The series «The human side of technology» will be presented on November 14th in Milan at the Catholic University (6 pm, Aula Magna Crypt, Largo Agostino Gemelli 1) in a meeting with Simone Natale, Giuseppe Riva and Massimo Sideri. The next releases on newsstands are: November 17, Simone Natale, Macchine ingannevoli. Communication, technology, artificial intelligence, with a preface by Don Luca Peyron; on 1 December, Giuseppe Riva, Social networks, with a preface by Andrea Gaggioli.
November 3, 2023 (modified November 3, 2023 | 8:40 pm)
© ALL RIGHTS RESERVED
#Journey #roots #artificial #intelligence #Corriere #time.news