A brief history of artificial intelligence embryo selection: from black-box to glass-box

With the exponential growth of computing power and accumulation of embryo image data in recent years, artificial intelligence (AI) is starting to be utilized in embryo selection in IVF. Amongst different AI technologies, machine learning (ML) has the potential to reduce operator-related subjectivity in embryo selection while saving labor time on this task. However, as modern deep learning (DL) techniques, a subcategory of ML, are increasingly used, its integrated black-box attracts growing concern owing to the well-recognized issues regarding lack of interpretability. Currently, there is a lack of randomized controlled trials to confirm the effectiveness of such black-box models. Recently, emerging evidence has shown underperformance of black-box models compared to the more interpretable traditional ML models in embryo selection. Meanwhile, glass-box AI, such as interpretable ML, is being increasingly promoted across a wide range of fields and is supported by its ethical advantages and technical feasibility. In this review, we propose a novel classification system for traditional and AI-driven systems from an embryology standpoint, defining different morphology-based selection approaches with an emphasis on subjectivity, explainability, and interpretability. Graphical Abstract A proposed classification system for artificial intelligence embryo selection models with different subjectivity, interpretability, and explainability.

Automated summary

With the advancement in computing power and accumulation of embryo image data, artificial intelligence (AI) has been applied in embryo selection in IVF. Machine learning (ML) has the potential to reduce subjectivity and save time, but the lack of interpretability in modern deep learning (DL) techniques has raised concerns. The effectiveness of black-box models lacks confirmation from randomized controlled trials, and recent evidence suggests underperformance compared to interpretable ML models. Interpretable AI is gaining support due to its ethical advantages and technical feasibility. A novel classification system for embryo selection models is proposed, focusing on subjectivity, interpretability, and explainability.