What is the mission of the Responsible Artificial Intelligence Task Force?
The goal is to create links with different actors and actresses of AI, who are not just academics talking to each other, or governments or AI professionals working in isolation.
The global partnership on AI (PMIA) – which has several working groups including that on responsible AI – was launched in 2020 at the end of a G7 initiative of Canada and France. There were initially 15 members, therefore 15 countries. We now have 25.
Together, we want to bridge the gap between practice and theory in AI, support and promote research in this sector, while aligning ourselves with priorities related to sustainable development.
You are taking over from Yoshua Bengio, the founder of Mila Quebec, as co-chair. How do you feel?
They are big shoes to fill, that’s for sure. I work a lot with Yoshua Bengio on different projects, so it will be the continuity of the projects that I already have with him.
His expertise is completely different from mine, he being a technical genius, me being more normative as a lawyer. But these approaches are complementary.
” I’m going to bring a different color, but not inconsistent with responsible AI development. »
I’m always proud when a woman manages to fit into the AI world. So to get into a position like this, where I can play a role not only as a leader of this organization, but also as a woman, bringing my different perspective, that makes me especially proud.
Is diversity and inclusion among your concerns?
This is at the heart of what I want to bring as co-chair. The group is already sensitive to these issues, and not just for women, but also for the countries of the South, which are often underrepresented.
” It’s a constant discussion in the group, that AI must be inclusive to be responsible. »
As a woman in a leadership position, I will be able to connect with other bodies and not skimp on messaging and accountability around diversity and inclusion. I already do, and I will continue to do so, because I find it very important.
Quebec know-how occupies a special place in this working group, why?
When the organization was created in 2020, it required a very significant investment of resources from France and Canada. As it evolves, we integrate other countries.
” We already have a great reputation in Quebec in terms of AI, and we have this concern as a leader to inspire others to join the movement. »
This desire of France and Canada to invest in a global partnership shows their ambitions to put the responsible development of AI at the heart of their priorities.
What are the achievements made so far?
In the fight against climate change, for example, the task force recently issued 50 recommendations to put AI to workHave (New window)Have.
A first proposal for a work card (road map) was presented at COP 26 and the PMIA summit to raise awareness among governments. A government interested in this would therefore already have the beginning of a useful guide for its action, but it is more in a voluntary form than it is at the moment.
How AI can help regulate judged online content
harmfulon social networks, for example?
We see the important effect that social networks have on democracy. There have been several examples of this in recent years. AI algorithms used in different social networks can lead Internet users in a certain direction and reinforce so-called echo chambers.
The working group is trying to develop a mechanism for assessing the factsHave (New window)Have (fact-finding) to better frame the content. We want to make sure we ask the right questions so that it doesn’t perpetuate potential abuse.
How is AI used in drug discovery to address diseases affecting poorer populations?
The idea is to ask how we can use AI to accelerate the development of certain drugs. This is a long process in which there are huge gains to be madeHave (New window)Have in time, and therefore in money. And this is very important for health systems.
What would you like to bring to the Responsible AI Working Group?
During my mandate, I will focus on four main priorities. The first is to maintain the work already started by the group on climate change, social networks and health.
But above all, as a lawyer, and also as a person, I have a strong concern for human rights in relation to AI. I would like to develop a new project in line with that.
We have a duty to ensure that AI respects human rights. It is a very strong legal and social pact, just like the Universal Declaration of Human Rights, on which there is consensus among States.
” Is AI discriminatory? Does it provide equitable access to health care? Does she respect privacy? These are extremely important concerns. We won’t get responsible AI if it negatively influences human rights. »
The PMIA is still a very young organization, and I would very much like to increase its impact and visibility by strengthening the links with governments. That way, when we issue recommendations, we will arouse their interest and their accountability so that they take action.
How do you foresee the co-presidency with Raja Chatila, from the Sorbonne University?
Raja Chatila has similar expertise to Yoshua Bengio, both coming from the more technical field of AI. It may be a less natural alliance with the human sciences, but it can be very fruitful. I can’t wait to see how we will manage to combine our complementary expertise to do something more ambitious.
This innovative partnership between the human sciences and AI must be maintained. It is in this way, with a better understanding of both areas, that we will succeed in developing responsible AI.