AI being used in IVF embryo selection
Media reports from India say AI tools are being used to analyse embryos and support selection decisions in IVF, with claims of improved success rates in local clinics. The coverage positions embryo assessment as a constrained, image‑heavy task where AI is being adopted ahead of messier parts of reproductive and maternal care. (indiatoday.in)
In vitro fertilization already relies on pictures: embryologists grade embryos in the lab, and clinics in India are now adding artificial intelligence to help rank them for transfer. (indiatoday.in) In a standard in vitro fertilization cycle, eggs are fertilized outside the body, embryos grow for several days, and a clinic chooses which embryo to transfer first. That choice has traditionally been based on morphology, meaning how the embryo looks under a microscope. (asrm.org) Artificial intelligence fits that step because embryo assessment is image-heavy and repetitive: software can score static blastocyst images or time-lapse videos for features linked to viability. A 2025 Nature Communications paper described a model trained on about 18 million embryo images for tasks including quality scoring, stage prediction and ploidy prediction from images alone. (nature.com) Indian clinics are now pitching those tools as part of routine care. India Today reported on April 11 that fertility centers in India are using artificial intelligence to analyze embryos and support selection decisions during in vitro fertilization. (indiatoday.in) At a New Delhi launch on April 1, Gaudium IVF introduced two tools from United Kingdom-based IVF 2.0: SiD for sperm identification and ERICA for embryo ranking. The Hindu reported that ERICA analyzes static blastocyst images and that Gaudium’s chairperson, Dr. Manika Khanna, said the tools review 2.5 million parameters and can raise first-cycle in vitro fertilization success rates by 5% to 7%. (thehindu.com) Those clinic claims sit alongside a more mixed research record. A 2024 multicenter randomized trial in Nature Medicine, covering 1,066 patients at 14 clinics in Australia and Europe, found a deep-learning embryo selection system did not prove non-inferior to standard morphology for clinical pregnancy rate. (nature.com) Other trials are still moving through the field. A 2025 Fertility and Sterility abstract described a prospective randomized trial testing whether artificial-intelligence embryo selection was non-inferior to traditional morphology for clinical pregnancy, and several newer protocols are still focused on proving benefit in live birth and pregnancy outcomes, not just prediction accuracy. (fertstert.org, pmc.ncbi.nlm.nih.gov) Professional guidance still treats embryo assessment as an uncertain process. The updated European Society of Human Reproduction and Embryology and ALPHA consensus said embryo evaluation remains essential but challenging, and the American Society for Reproductive Medicine said new tools, including artificial intelligence, are being developed alongside older morphology-based grading. (rbmojournal.com, asrm.org) India is a large market for that experiment. Government-backed law now regulates assisted reproductive technology clinics and banks under the Assisted Reproductive Technology (Regulation) Act, 2021, and industry reports estimate India performs roughly 200,000 to 300,000 in vitro fertilization cycles a year. (icmr.gov.in, enira.co.in) The demand side is also clear: the World Health Organization said in 2023 that about 17.5% of adults worldwide, roughly 1 in 6, experience infertility. In that setting, embryo ranking is becoming one of the first parts of fertility care where artificial intelligence is moving from research papers into everyday clinic workflows. (who.int, indiatoday.in)
