AAA Mini-Research Series – Part 1
- Introduction
In vitro fertilization (IVF) has undergone significant advancements over the past few decades, with automation and artificial intelligence (AI) playing an increasingly central role in laboratory procedures. These technologies are not only enhancing efficiency and precision in IVF labs but are also reshaping the way embryologists assess, select, and monitor embryos, sperm, and overall laboratory workflows. However, while proponents argue that AI and automation increase success rates, reduce errors, and standardize processes, critics highlight concerns regarding ethical considerations, AI bias, and over-reliance on machine learning models.
This article delves into the latest advancements in AI-driven embryo selection, sperm sorting, time-lapse monitoring, and workflow automation, presenting trending facts, major laboratory technologies, and equipment innovations. As the first part of our series on “Innovations in IVF Lab Equipment,” this discussion will critically examine the latest studies, technological breakthroughs, and potential challenges in integrating AI and automation into fertility treatments.
2. AI in Embryo Selection: Enhancing Precision in IVF Outcomes
Embryo selection is one of the most crucial steps in IVF, traditionally performed by embryologists using morphological assessments and preimplantation genetic testing (PGT). However, AI-powered deep learning models have emerged as a non-invasive and automated alternative for embryo grading.
Duzcu et al. (2024) found that AI-assisted embryo selection can automate time-lapse imaging analysis to identify optimal embryos with a higher implantation probability compared to traditional grading. Their study highlights the use of deep learning models in improving embryo viability assessments and reducing subjectivity in selection. Paya Bosch (2024) focused on deep learning for embryo selection, demonstrating how AI-driven automation improves the consistency of grading compared to human embryologists. The study introduced STORK-AI, a deep learning model for non-invasive embryo selection.
While AI models like STORK-AI and Life Whisperer have shown impressive accuracy in embryo selection, some experts caution against over-reliance on algorithms, arguing that AI lacks human intuition and may misinterpret ambiguous morphological features. There are also concerns regarding bias in training datasets, potentially favoring embryos with specific morphological characteristics.
3. AI in Sperm Selection: A Paradigm Shift in Assisted Reproduction
Sperm selection has traditionally relied on manual microscopy techniques, such as Intracytoplasmic Morphologically Selected Sperm Injection (IMSI) and Intracytoplasmic Sperm Injection (ICSI). However, AI-driven computer vision and microfluidic sperm sorting technologies are transforming how sperm are selected for fertilization.
Montjean et al. (2024) tested SiD (Sperm Identification Device) and concluded that AI-driven sperm selection performed similarly to human embryologists but with faster and more standardized results. Hozyen et al. (2024) examined the role of AI-enhanced sperm selection in IVF and found that automated sperm sorting improved fertilization rates and embryo quality. The study highlighted the use of SpermVision and CASA-based AI models in selecting motile, DNA-intact sperm for ICSI.
Critics argue that AI-driven sperm selection systems may not yet fully replace embryologists, as factors beyond morphology and motility (e.g., sperm DNA integrity and oxidative stress) also influence fertilization success. Moreover, commercial AI-based systems remain expensive, raising concerns about accessibility and cost-effectiveness.
4. Time-Lapse Monitoring & AI Integration
Time-lapse monitoring (TLM) has become a staple in modern IVF labs, allowing continuous embryo observation without disturbing culture conditions. When integrated with AI, TLM automates embryo scoring, tracks developmental patterns, and predicts implantation potential.
Janmohamed et al. (2024) examined AI’s ability to predict blastocyst formation and implantation potential by analyzing time-lapse morphokinetic markers. Zhylko et al. (2024) introduced an AI-based embryo tracking system that pinpoints the optimal transfer day, improving implantation rates.
While TLM with AI integration improves precision and consistency, critics argue that there is no substantial clinical evidence that it significantly increases live birth rates over traditional selection methods. Additionally, cost remains a barrier, with high-end incubators and AI-powered systems increasing treatment expenses.
5. Ethical & Regulatory Considerations
With AI playing an increasingly autonomous role in IVF, ethical concerns, biases, and regulatory challenges have become central to the discussion. Key concerns include:
- Bias in AI Training Data: Some AI models may favor embryos based on pre-programmed datasets, leading to potential unintended selection biases.
- Lack of Human Oversight: Over-reliance on AI could reduce the role of embryologists, raising ethical concerns regarding autonomy and machine-driven decision-making.
- Regulatory Challenges: Organizations such as the European Society of Human Reproduction and Embryology (ESHRE) and the U.S. Food and Drug Administration (FDA) are still evaluating how AI-powered IVF systems should be standardized and regulated.
De Proost & Segers (2024) proposed a principle-based framework for regulating AI in IVF, arguing that AI should complement, but not replace, human expertise.
Conclusion
AI and automation are redefining the IVF laboratory, offering higher precision, efficiency, and standardization. While AI-driven embryo selection, sperm sorting, and time-lapse monitoring are promising innovations, they also introduce challenges related to ethics, cost, and human oversight. The future of IVF will likely see further AI integration, but ensuring responsible and well-regulated use will be critical.
Would AI fully replace embryologists in the future? The debate continues.
Photo Credit: Onyema Priscilla Uloma, Viva Fertility Clinic Abuja
References
- Cohen, J., Costa-Borges, N., & Mendizabal, G. (2024). Far away so close: Remote ICSI using an AI-powered robot. Human Reproduction. Retrieved from https://academic.oup.com/humrep/article-pdf/doi/10.1093/humrep/deae108.218/58509953/deae108.218.pdf
- Duzcu, T., Ergun, Y., & Basar, M. (2024). Artificial intelligence in IVF laboratories: Elevating outcomes through precision and efficiency. Biology, MDPI. Retrieved from https://www.mdpi.com/2079-7737/13/12/988
- Gaspard, L. W. C., Leisinger, C., et al. (2024). Does a dry or humid incubation environment affect continuous and undisturbed culture media protocols? Human Reproduction. Retrieved from https://academic.oup.com/humrep/article-pdf/doi/10.1093/humrep/deae108.589/58399438/deae108.542.pdf
- Hozyen, M., Sayed Azzouz, Y., Zaki, H. H., et al. (2024). Investigating the effect of AI-tool for sperm selection in IVF laboratories. Fertility and Sterility. Retrieved from https://www.fertstert.org/article/S0015-0282(24)01709-6/pdf
- Janmohamed, A., Nayot, D., Miller, R., et al. (2024). Artificial intelligence and oocyte/embryo assessment in cryopreservation cycles. Springer. Retrieved from https://link.springer.com/chapter/10.1007/978-3-031-58214-1_23
- Montjean, D., Pacios, C., Hamiche, G., et al. (2024). Automated sperm selection software (SiD) provides similar results as human selection for ICSI. Human Reproduction. Retrieved from https://academic.oup.com/humrep/article-abstract/39/Supplement_1/deae108.482/7703726
- Paya Bosch, E. (2024). Deep learning for the automation of embryo selection in an in vitro fertilization laboratory. Riunet UPV. Retrieved from https://riunet.upv.es/… 2/2/2025
- Zhylko, D., Pardo, S., Del Gallego, R., et al. (2024). A cutting-edge artificial intelligence system tracking embryo development and pinpointing the optimal day for blastocyst utilization. Human Reproduction. Retrieved from https://academic.oup.com/humrep/article-… 12/2/2025