Comment suivre tout ce que l’on lit sur l’IA en science ?
En santé, on voit de tout pour nous aider à réfléchir. Je ne cite que des articles récents au hasard :
Un bel éditorial (juillet 2024) du rédacteur en chef du BMJ « À la gloire de l’IA ennuyeuse » (image ci-contre) pour commenter une série d’articles dont celui-ci ‘Forget about replacing doctors with AI — just get our computers to work‘. Un coup de gueule court dont le titre est percutant !
Dans le JAMA Internal Medicine (juillet et août 2024), des questions : ‘Can artificial intelligence speak for incapacitated patients at the end of life?‘. Quand vous discuterez avec des enfants pour prendre une décision à propos d’un parent en réanimation, pourquoi ne pas utiliser toutes les anciennes visites enregistrées pour savoir ce que ce comateux (sans directives anticipées) aurait répondu ? Et un autre ‘How should medicare pay for artificial intelligence?‘ Allez-vous solliciter les payeurs du système de santé pour les coûts liés à l’IA, etc….
Comment exploiter le potentiel de l’IA de manière responsable ? Quatre experts répondent
Four scientists spoke to Nature about How to harness the power of artificial intelligence (AI) in research responsibly and ethically. J’ai repris la newsletter de Nature présentant l’article :
- Harness AI for good, keep ethical standards high
AI is just a tool — it’s up to its users to consider how it might reinforce discrimination, says computer scientist Ross King (UK, Sweeden), who helped create the Stockholm Declaration on AI for Science. “Already, it is a bit too late,” he warns.
- Understand the limitations of AI tools
Don’t try to use AI to solve every problem, says computer scientist Suresh Venkatasubramanian USA), who helped to co-author the first US blueprint for an AI Bill of Rights. “Right now, we are in the hype cycle, in which no one’s talking about the limits of these tools,” he says.
- Effective, ethical AI requires representative data
Computer scientist Nyalleng Moorosi (Lesotho) has a “passion for data representation: that is, who is included in the data sets and how they are portrayed”. She co-founded the non-profit organization Deep Learning Indaba so that “Africans will be not only observers and receivers of advances in AI, but also active shapers and owners of those advances.”
- Prevent AI-driven colonization in Africa
“I am concerned that Africa is being managed, against its will, as the least rewarded link in the global chain of the AI economy” as a source of training data or low-paid workers, says computer scientist Seydina Ndiaye (Senegal), who has co-founded several information-technology companies on the continent. He calls for African-led AI development and for researchers globally to consider how the Africans who contributed data for AI models get to benefit from those systems..
