An incomplete optimization: IVF directs AI at embryo laboratory while treatment discontinuation goes unmeasured

Reproductive medicine has invested four decades in optimizing the IVF laboratory; artificial intelligence (AI) applied to prediction, selection and automation represents the latest iteration of that investment. Meanwhile, treatment discontinuation, affecting 36% of patients, has received no systematic attention, despite its demonstrably larger effect on outcomes. A meta-analysis shows patients who complete three assisted reproductive technology (ART) cycles achieve a cumulative live birth rate of 57.9%; this becomes 42.7% when early discontinuation is included. In New Zealand, active patient management in a funded programme reduced intercycle dropout to 10%, and a randomized trial showed that structured psychological support reduced discontinuation by 64%. However, the 2025 International Glossary on Infertility and Fertility Care - 348 consensus definitions endorsed by 11 organizations - contains no entry for patient-level treatment discontinuation. This Countercurrent argues that IVF has concentrated innovation in the embryo laboratory while neglecting an outcome of comparable magnitude, namely that AI could be redirected to identify patients at risk of dropout; and that a four-step agenda - standardized definition, institutional measurement, quality reporting, and patient-facing disclosure - is achievable with existing clinic data.