Ahead of the (Robocup 2023)[https://2023.robocup.org/en] (which will take place in Bordeaux from July 4th to 10th), I had the pleasure of answering questions from Maxime Giraudeau, a journalist at La Tribune (Sud-Ouest edition), regarding the hopes and limitations surrounding AI. The interview mainly focused on the scientific and societal issues related to the Trustworthy AI Chair that I hold with the support of the Bordeaux University Foundation. We discussed various aspects of AI and its potential uses.
Once again, I emphasized that even though the latest accessible tools may give the illusion of a deep understanding of certain subjects, it is absolutely not the case, and errors always occur. The concept of truth does not exist. The transition to the real world, and the meaning it carries, is inaccessible to these generative systems. Therefore, it is crucial to understand their limitations before expecting any benefits from them, or to restrict their use to light (without real consequences) or human-supervised applications. The question of responsibility is equally important and remains unanswered.
These are fascinating topics and a challenge for the journalist I spoke with: summarizing a discussion that took off in different directions!
(You may find the article here (behind a paywal))[https://objectifaquitaine.latribune.fr/innovation/2023-07-07/comment-redonner-confiance-en-l-intelligence-artificielle-968465.html]