Definition
A stakeholder is an actor (an individual or a group), who is directly or indirectly concerned by a decision or a project, because of the positive or negative impact it may have on them (« Stakeholder (Corporate) », 2024).
Context
We have begun a partnership with two researchers, which will run through 2024. Over the next few months, our team will be taking part in reflexive workshops. The aim is to question and train our exercise of responsibility as designers of translation tools using automatic learning techniques (known as “AI”): what impacts can our activity have, and how can we avoid deleterious effects on our stakeholders?
(Article – in French – to find out more about this objective)
This blog post is about our first face-to-face workshop, which follows an introductory session where we were made aware of the specific AI issues requiring our vigilance.
The scope of our stakeholders
The session was divided into two activities, led by the researchers who were accompanied by two interns. The first activity aimed to reflect in depth on the groups of people impacted by Xenizo’s activity: our stakeholders.
To do this, we used a responsible design tool, the “inclusive panda”, which we filled in together. We had to answer questions as simple as “who is our main target?”, to much less obvious ones like “which people oppose what we do, to the point of using our productions against us?”. In between these two extremes, this exercise has uncovered anyone who has anything to do with our activities, from near or far.
Thanks to this tool, our team has had the space to unfold thoughts that it only skims over the rest of the time. On a day-to-day basis, it doesn’t seem useful to ask ourselves which people are not impacted by our activities because they decide they’re not interested.
Taking the time to reflect, together, on all the people impacted directly or indirectly by our work forces us to accurately assess our responsibility when designing tools for translators of biblical resources.
What could possibly go wrong?
For the second activity, we imagined disaster scenarios in small groups. Playing the role of journalists, we had to produce a newspaper front page about an ethical-responsible disaster involving Xenizo: what went wrong, and what were the consequences?
Whatever the group, the scenario led to an outpouring of violence.
So, beyond the creative aspect of the workshop, the discussions held revealed to us the potential impact of our work in the event of a problem. By reconnecting these prospective catastrophic scenarios to the present time, we targeted areas of risk in what we do, and started thinking about solutions.
What next?
This workshop was the first in a series, and the conclusions of this work will gradually take shape. One of the objectives of this partnership is to write an AI ethical charter for Xenizo, which will be the result of our reflections on the subject, and which will guide design decisions for biblical resources translation tools within the company.
We’ve started here by thinking about the people impacted by our activities and the extent of the damage that could be done if something goes wrong. I imagine that, over time, we’ll come up with a framework for solutions. We’ll find out in the following workshops.
