Artificial intelligence in health seen from the year 2044

LThe following text is fiction. It aims to raise awareness and question expectations regarding the latest fashionable technology: artificial intelligence (AI). It echoes a column published in The Worldon May 21, 2024, which takes stock of twenty years of dazzling development in genetics in health, and proposes to imagine, in twenty years, the same questions applied to AI.

It’s 2044. In twenty-five years, artificial intelligence has arrived everywhere. Its use has become widespread in health. Much was expected of it, people complain that European states missed the boat, and yet its impacts are still poorly understood. The French Federation of Artificial Intelligence in Health concludes eighteen months of work by the various learned societies using AI in health. Here is its report.

“In assisted consultation meeting”

“Artificial intelligence seems to impose its definitive and irreversible accuracy on everyone. Even in everyday language, we claim an identity defined by it: “this is what ChatGPT says about me”, an attitude that would be a “déjà vu by AI”. However, uncertainty in AI is not only omnipresent but above all it is plural. Professionals in the field know this well: we do not necessarily know everything, we do not even know if we could know everything and, moreover, we do not always know if our knowledge applies to a given situation.

Thanks to the deployment of new computing techniques and better availability of data, including synthetic data, AI in medicine has undoubtedly been able to provide assistance to patients and their families, both from the point of view of diagnosing a pathology, providing AI-assisted advice, establishing a predisposition to a disease in AI, and for a therapeutic choice, or even the discovery of a therapy.

In “oncoIA”, faced with an accumulation of cancers occurring most often at unusually early ages in families, it is now possible to assess cancer risks. This opens the way to screening pathways, or even preventive actions, including surgeries, for families, in the event of a quantified predisposition in AI. (…)

Also read the article: Article reserved for our subscribers Artificial Intelligence: “We must make health data a common good for research”

At the same time, in oncology, the characterization by AI of the anomalies of a tumor can be decisive for the choice of the best treatment, particularly for aggressive forms of cancer or relapses.

You have 61.79% of this article left to read. The rest is reserved for subscribers.

source site-27