Témoignage UTCéen - AI for Health - An opportunity for the AI healthcare ecosystem

AI for Health - An opportunity for the AI healthcare ecosystem

This year's edition of “AI for Health '' conference took place on Nov 16th 2022 in Station F, in the heart of the Parisian startup ecosystem. 

The event was divided into three main parts:

•  A standard exhibition floor, where startups and companies alike presented their products and services
• A main floor, where key stakeholders, researchers and people from industry discussed topics related to data and AI for health
• A secondary floor, where startups were invited to pitch their ideas for a duration of about 7 minutes

AI in healthcare has been a growing source of innovation, starting from a paradigm of belief networks and logical AI to assist diagnosis, via rule based methods and knowledge databases. It then evolved, together with the arrival of deep learning, to image recognition, aiming to identify lesions in medical imagery. We are now witnessing a new cycle for AI applied to healthcare, mainly coming from graphs and multimodal AI. The most famous approach of this trend is protein folding for drug discovery.

One interesting trend I witnessed while at this conference is common to most of the field of AI. Shifting from trying to do better than the professional, towards a "hand in hand" approach. The industry has realised that the biggest potential in AI is to use it in combination with the medical experts, trying to empower them rather than to replace them. Logically, recently a large proportion of startups applying AI to healthcare therefore have foundres with a medical background.

In parallel, France is a quite active area in the field of AI, thanks to the many great universities and research institutes that are present. Taking advantage of that position, we are leading an innovative centralization of medical big data with the Health Data Hub since 2019. This aims to build a line of direction for data governance and its uses, at an European level.

Many of the stakeholders of those European projects were present, and it was an exciting opportunity to exchange and discuss the future of AI in this niche domain, which is quite different from the rest. The key here lies in building a safe way to handle this confidential and sensitive data, in a way that makes it structured enough for machines to understand.

The road is still long to leverage all the possibilities that exist in this exciting space, and it is great to see that these kinds of events are here to help start the communication that is needed between the actors. One group was however notably absent from the event: patients. It could be an improvement for future editions to try and bring them closer to these discussions.

 Overall, it was a very interesting event, worth attending, and I want to thank the UTC alumni association for giving me the opportunity to attend for free.