A software engineer goes to ESHRE...

As a software engineer, I recently had the opportunity to step into the realm of medical conferences. The first-time experience at the European Society of Human Reproduction and Embryology (ESHRE) Congress, where I was honored to be invited as a speaker, is a steep departure from the typical academic or tech industry events that I’m accustomed to.

Although branded European, the event had a global reach with more than 10000 participants coming from all corners of the globe. I had the privilege to contribute my talk about the certification of artificial intelligence to the pre-congress course on Artificial Intelligence in Assisted Reproductive Technology (ART).

The course program offered a balanced blend of technical and domain-specific discussions. As anticipated in an AI-focused event, the technical sessions delved into algorithms and regulatory procedures associated with medical devices and software. However, it was the ART-related talks that truly caught me by surprise. Despite the abundance of medical terminology and copious amounts of data, which presented a challenge to comprehend fully, I was captivated by the profound mastery of the mathematical underpinnings that form the foundation of AI technology.

The speakers, primarily professionals from reproductive clinics, exhibited a remarkable expertise within their domain. However, their fluency in the AI jargon commonly used in the tech industry was limited. Consequently, the latest advancements in generative AI, particularly large language models (LLMs), appeared intimidating to them. The incorporation of AI-related technologies within clinical settings posed a significant challenge, particularly when considering the complex regulatory landscape. Surprisingly, for someone familiar with regulatory intricacies, the field of human reproduction presents even greater complexities in terms of both regulations and ethical considerations. Adding to the complexity, these issues varied substantially across jurisdictions, imposing significant limitations on practitioners and affecting their ability to collaborate on the development of novel technical solutions.

The course program concluded with an engaging panel discussion that fostered active participation from both the presenters and the audience. The conversation revolved around several overarching themes, such as the perceived gap between practitioners and technology within the field of reproductive medicine, and strategies to enhance cooperation. A notable challenge highlighted was the absence of shared reference databases, which hinders dissemination of best practices. Similarly, interoperability issues between tools lead to fragmented workflows. However, the potential solution of leveraging the Fast Healthcare Interoperability Resources (FHIR) standard was proposed as a means to address this challenge. Perhaps the most significant point of contention focused on the ethical considerations surrounding the use of AI in human reproduction. Ultimately, the consensus was reached that it is ethically unjustifiable to not harness the power of AI in this domain, emphasizing the imperative to integrate AI technologies responsibly and ethically.

Thanks Sara Zdravka Veleva and José Alarcón Roldán for inviting me to this great event! I surely enjoyed it!