27B - Artificial Intelligence Is Useful for Embryo Selection in IVF

Summary

The use of Artificial Intelligence (AI) aims to revolutionise the way we work within the clinical IVF laboratory by offering invaluable assistance to the scientists selecting embryos for transfer. AI does this by removing the human operator from the process, allowing digital systems to feed on the data through a process called datamining. The AI system establishes artificial neural networks and essentially learns which embryo parameters are associated with desirable or undesirable outcomes (1). The aim is to accurately detect the viability of embryos and rank them according to their implantation potential, thus increasing confidence in single embryo transfer and time to pregnancy. By performing entirely objectively, with precise scrutiny and with a wealth of data not accessible to the scientist, this computerised approach is designed to outperform the embryologist. However, despite efforts to establish a firm place for AI during routine IVF, to date there is little evidence it will serve as a useful clinical tool. Although the non-invasive nature of AI appeals to professionals and patients as being seemingly harmless, the technology has multiple flaws that currently prevent its promotion to anything other than yet another useless and risky IVF ‘add-on’.