As the previous posts indicate, I keep a skeptical eye on all the Artificial Intelligence hype. But only criticising doesn’t suit my style. That’s why I want to take action. Below is a first draft towards this goal.
A confusion of tongues around “AI” 👅
Recently, I came across an article and reflected on how it could pass the peer review process. The authors tried bringing out a new concept, “Cognitive computing” or “Cognitive Informatics”, described as an “advancement” in Artificial Intelligence. Genuinely, they were just recycling old ideas with a new name (either that or I’m skipping the next technological revolution). Hither their definition:
“A trans-disciplinary investigation of computer science, information science, cognitive science, and intelligence science that investigates the brain and natural intelligence's internal information processing mechanisms and processes, as well as their cognitive computing engineering applications”.
It looks curiously familiar with the initial aim of AI and its presentation in the associated Wikipedia page… 🤔 This looks less like a particular mistake in the peer review process than a confusion of tongues surrounding AI. The name itself is misleading. On this point, Alan Turing, one of the creators of AI, disclaimed the literal meaning of Artificial Intelligence (as replicating human intelligence). Rather, he drew the parallel: a “submarine that swims” represents a figure of speech; nobody expects it to actually swim like a fish.
This article reveals two points:
The peer review process is flawed, but not every system is foolproof.
There is a lack of understanding about AI…
With my engineering studies, I developed a technical baggage to comprehend the key concepts in AI. I'm therefore at the border between two worlds: the “AI” and the “management sciences” communities. They are often disconnected and little they know about each other.
Hence, as part of my PhD topic, I wish to clarify and explain the workings behind AI to the “management sciences” community. Likewise, I hope my findings will help in managing AI projects more rationally and responsibly 🤞🏻
But… What does it mean to translate? 👽
This line of thoughts leads to imagine myself as a translator. But it is not like translating from English to French, although traditional translators have to include -or at least consider- cultural elements. In our case, it might mean to translate concepts form a scientific paradigm to another… Remember my previous article? (If it looks like the order of my posts is prepared, rest assured: it is totally fortuitous! I’m lost. I don’t know where I go. More on one of my older posts: “The fog in a PhD” 😉)
I think Michel Callon could assis us. He is a French researcher in Science and Techniques Studies (STS), a branch of sociology. In 1986, he wrote an article titled: “ELEMENTS FOR A SOCIOLOGY OF TRANSLATION: The domestication of scallops and fishermen in the Bay of Saint-Brieuc”1. This article serves the purpose to alter the way STS was done by proposing a new method to analyse social phenomena. Hence the “elements for a sociology of translation”.
He narrates how three scientists tried to “domesticate” both the scallops and the fishermen. The popularity of scallops lead to overfishing, which in turn reduces their population, increasing the price, making it more alluring commercially, and so on. The idea was to farm those invertebrates, instead of just fishing savage ones. However, this implied to also redefine the work of fishermen.
Thanks to this example, Callon labels “translation” the process of:
Establishing a problem, like how the three scientists did. It is not only about generating questions and bringing out a problematic situation. It means defining which actors are implied and the relationships needed for a solution to be found.
Incentive arrangements come next. The defined actors (the scallops and the fishermen) are not obliged to accept how their identity is defined. They could perceive themselves differently. That is why, an incentive needs to be found so they accept the definition established in the problematisation.
Next, there is the enrolment, the mechanism by which the actors accept (or not) the relationship that links them to the others. It isn’t a binary decision, but instead it implies a series of negotiations and mutual adjustments between the actors.
But are the spokesperson representative? Who represents who? There is no direct or comprehensive contact with the group behind the concept of actors. Only a limited number of individuals from each party are in negotiations. Do they really represent the whole crowd behind them?
If this process goes well the traduction is a success 🎉 There are displacements of interests, of aims, of devices, etc. as they are inscribed in a reality. However, before becoming a result, the traduction is a process in which treason is possible. The definition of the problem, the incentives, the enrolment or the spokesperson might betray the “situation”. As a consequence, the situation doesn’t evolve as planned.
Illustration with one of my current streams of work 📝
Going back to my role in translating AI concepts to the management sciences community, I have to better define the problem: the actors might not be as black and white as “AI guys” vs “organisation researchers”, their relationships could take various configurations and I have to seek representative spokespersons.
To illustrate it with my ongoing paper project: I need data scientists to describe an AI tool they have developed. This why, my theoretical framework could be tested by analysing those projects. Concluding results could become an article!
But, to look for those data scientists means: I have to find a representative target group that identifies itself with my problem and that is ok to be enrolled in the project. Likewise for the community to which I will submit my parer (who is represented by the reviewers that might read my article).
Takeaways 🎁
Were the “elements for a sociology of translation” clear enough? Take it as a possible framework to analyse and generate insights on a project you are doing.
A more skeptic view towards what you read about AI. There is much hype going on, and many try to profit from the zones of uncertainty. (Actually, some organisational theories were developed around how individuals take profit from uncertainty and gain advantage out of it.)
There is a small peek at my current research. I seldom speak about it… I’m both timid and advancing very slowly hahaha🙈
Well, thank you for reading. I hope you enjoyed this post 😁 Don’t hesitate to leave a comment in the section below or discuss directly with me! Any feedback is welcome and appreciated 🙏🏼
The article is in French, and the original title is: “ÉLÉMENTS POUR UNE SOCIOLOGIE DE LA TRADUCTION: La domestication des coquilles Saint-Jacques et des marins-pêcheurs dans la baie de Saint-Brieuc”. To read the article: click here.