AI in Fertility: How Algorithms Are Choosing Embryos

Why Embryo Selection Matters So Much

One of the most critical moments in the entire IVF process happens in the lab, not the exam room: deciding which embryo to transfer.

The average IVF success rate is around 30% per cycle. That means roughly 7 out of 10 transfers don't result in a pregnancy. While many factors contribute to this — uterine receptivity, immune responses, hormonal balance — embryo quality is one of the biggest variables clinics can influence.

Traditionally, embryologists assess embryo quality by looking through a microscope. They evaluate factors like cell symmetry, the pace of cell division, the appearance of the inner cell mass (the part that becomes the baby), and the trophectoderm (the part that becomes the placenta). Based on this visual inspection, embryos are graded — typically on a scale like AA, AB, BA, BB — and the highest-graded one is selected for transfer.

The problem? This process is subjective. Two experienced embryologists looking at the same embryo can assign different grades. One study found inter-observer agreement on embryo grading was as low as 60-70%. That level of inconsistency, in a process that costs $15,000–$25,000 per cycle, is a real problem.

How AI Embryo Selection Actually Works

AI-based embryo assessment isn't a robot picking embryos out of a dish. It's software — trained on thousands (sometimes tens of thousands) of embryo images — that analyzes visual data and outputs a score or ranking.

Here's the general process:

1. Time-lapse imaging captures continuous data. Modern IVF labs use incubators equipped with cameras (called time-lapse monitoring systems) that photograph each embryo every 5-20 minutes from fertilization through Day 5 or 6. This creates a detailed developmental "movie" of each embryo — something no human could observe continuously.
2. The AI analyzes thousands of data points. The algorithm examines morphological features (shape, size, symmetry), morphokinetic patterns (the timing and speed of cell divisions), and developmental milestones. It compares these against patterns from its training data — embryos whose outcomes (implantation, pregnancy, live birth) are already known.
3. The system produces a score or ranking. Each embryo receives a numerical score representing its estimated implantation potential. Embryologists then use this score alongside their own assessment and the patient's clinical context to make a transfer decision.

The Major AI Tools in Use Today

Several AI platforms are now commercially available or in advanced clinical testing. Here are the ones you're most likely to encounter:

What the Research Actually Shows

This is where things get nuanced — and where you need the full picture before making decisions about your care.

The Promising Findings

Retrospective studies (looking back at data from past IVF cycles) have been encouraging. AI systems consistently demonstrate high accuracy in grading embryo quality, often matching or slightly exceeding the performance of experienced embryologists. The STORK system, for example, achieved an AUC of greater than 0.98 when predicting blastocyst quality, and it generalized well across clinics in different countries.

ERICA showed it could predict whether an embryo was chromosomally normal (euploid) from a single photograph — something that typically requires an invasive biopsy and $4,000–$5,000 in genetic testing (PGT-A). If AI could reliably replicate that non-invasively, it would be transformative.

The Sobering Reality Check

The most rigorous test of AI embryo selection to date came from a large randomized controlled trial (RCT) published in Nature Medicine in 2024. This multicenter study across 14 IVF clinics in Australia and Europe compared iDAScore (the AI) against standard embryologist assessment.

The results: iDAScore achieved a clinical pregnancy rate of 46.5%, compared to 48.2% for traditional morphology-based selection. The study was unable to demonstrate that AI was even non-inferior to experienced embryologists — meaning it couldn't prove AI was "at least as good as" human assessment.

Non-Invasive Ploidy Assessment: The Next Frontier

Perhaps the most exciting application of AI in embryology isn't just grading how an embryo looks — it's predicting what's happening inside its chromosomes.

Preimplantation genetic testing for aneuploidy (PGT-A) is currently the gold standard for assessing whether an embryo has the right number of chromosomes. But PGT-A requires biopsying a few cells from the embryo — an invasive procedure that adds cost and, while generally considered safe, introduces a small element of risk.

AI tools like STORK-A are now being developed specifically to predict ploidy status from time-lapse images alone — no biopsy needed. In testing, STORK-A predicted aneuploidy versus euploidy with an accuracy of about 69% using images combined with maternal age and morphokinetic data. For complex aneuploidy specifically, accuracy climbed to approximately 78%.

That's not high enough to replace PGT-A yet. But for patients who can't afford the $4,000–$5,000 for genetic testing, or whose embryos aren't suitable for biopsy, AI-based ploidy estimation could provide meaningful guidance that wouldn't otherwise exist.

The Advantages and Limitations

The Regulatory Landscape

If you're wondering "who's making sure this is safe?" — it's a fair question, and the answer is: it's complicated.

In Europe, iDAScore has received CE mark certification, meaning it has met European standards for safety and performance as a medical device. In the United States, the FDA regulates AI tools in fertility when they qualify as Software as a Medical Device (SaMD). However, many embryo-ranking tools currently operate under the category of "clinical decision support," which has a lighter regulatory touch.

The AI Fertility Society (AIFS) — a growing professional organization — is actively working to establish ethical frameworks and best practices. Their 2026 conference in Cascais, Portugal, will focus specifically on responsible AI implementation in reproductive medicine.

What this means for you as a patient: ask your clinic specifically which AI tool they use, what regulatory status it has, and how much weight the embryology team gives to AI scores versus their own assessment. A transparent clinic will be happy to explain.

What to Ask Your Fertility Clinic

If you're going through IVF (or considering it), here are the right questions to raise about AI in your clinic's lab:

1. "Do you use any AI or machine learning tools for embryo assessment?" — Not all clinics do, and some use time-lapse monitoring without AI scoring.
2. "Which system do you use, and is it commercially validated?" — There's a difference between a published, validated tool and something a lab developed in-house.
3. "How does the AI score factor into your transfer decisions?" — You want to know if it's one input among many or the primary driver.
4. "Does using AI change the cost of my cycle?" — Most clinics absorb the cost, but some charge additional fees for time-lapse monitoring or AI-enhanced assessment.
5. "Would AI change your recommendation about PGT-A?" — For patients on the fence about genetic testing, this is worth asking.

The Ethical Questions Worth Thinking About

AI in embryo selection raises legitimate ethical questions that the fertility community is actively debating:

Equity and access. AI tools are expensive to develop and integrate. Will they primarily benefit patients at well-funded urban clinics, widening the gap between high-resource and low-resource fertility care? Cloud-based AI platforms could help by bringing sophisticated analysis to smaller clinics — but only if the economics work out.

Transparency. When an algorithm helps decide which embryo gets transferred (and by extension, which potential person comes into existence), patients deserve to understand how that decision was made. The "black box" nature of deep learning models makes this challenging.

Data privacy. AI systems require vast amounts of embryo data for training. Whose embryo images are in these datasets? Were patients adequately informed about how their data would be used? As these tools proliferate, data governance becomes critical.

Scope creep. Current AI tools focus on predicting pregnancy viability — not selecting for traits. But as technology advances, the line between "selecting the healthiest embryo" and "selecting for preferred characteristics" could blur. The fertility community, regulators, and society need to set clear boundaries now.

The Bottom Line for Patients

AI in embryo selection is real, it's happening now, and it's getting better. But it's not magic — and it doesn't replace the skill, experience, and clinical judgment of a good embryology team.

Here's the honest assessment:

• If your clinic uses AI-assisted embryo assessment, that's a reasonable indicator that they're investing in modern laboratory technology.
• AI may be most valuable for patients with difficult cases — those who've had multiple failed transfers, borderline embryos, or limited numbers of embryos where every selection decision counts.
• Don't choose a clinic based solely on whether they use AI. Success rates, embryologist experience, lab quality, and your comfort with the medical team all matter more.
• The technology will likely improve significantly over the next few years. What's inconclusive now may become standard of care by 2028-2030.

The future of embryo selection will almost certainly involve some form of AI. The question isn't whether it will happen, but how quickly the evidence catches up with the promise — and whether patients can access it equitably when it does.

Preparing for IVF: What You Can Control

While you can't control which AI tool your lab uses, you can optimize the factors within your reach. The quality of your eggs and sperm going into the process matters enormously — and there's strong evidence that lifestyle and supplementation in the 2-3 months before a cycle can make a difference.

Our picks for IVF preparation:

• CoQ10 (Ubiquinol) — Research supports 400-600mg daily for egg quality, particularly for women over 35
• It Starts with the Egg by Rebecca Fett — The definitive guide to evidence-based IVF preparation
• High-quality prenatal vitamin with methylfolate — Start at least 3 months before your cycle