Psst… We received a tip…

A couple in an IVF clinic are trying a new AI test. On the monitor are two day-5 embryo photos from their cycle. The new AI test can tell them which embryo has a stronger physical composition to survive the next steps. 

The first photo shows a strong-looking embryo, with clean edges, even shape, developing right on schedule. The second embryo looks a bit behind and gets a lower score. 

Later, the genetic testing results arrive. The embryo that was just scored as more “strong” actually shows a chromosome problem that often blocks a healthy pregnancy or ends early. The plainer one is chromosomally normal. Two signals pulling in different directions. 

There is tension in this testing. This week's patent uses image scoring as a fast read on physical viability from a single photo. However, this is different from the controversial genetic testing that reads the DNA code. In a world concerned about "designer babies", this week's newsletter will explore this patent as a useful alternative viability test, while also exploring perspectives on its shortcomings and its tension with genetic testing that spark heated moral debates.

This week’s patent scores embryos using a day 5 photo. The score is a read on physical strength in the near term, not a peek at DNA. Clinics like it because they already take these photos. You get a quick signal before making a high stakes call. 

ASTEC filed the patent, which is a 45-year-old, ISO 13485–certified Japanese maker of IVF lab gear. The inventors are associated with Presagen/Life Whisperer which is the Adelaide AI outfit behind embryo-scoring software. “Life Whisperer” produces technology for both morphological testing (as in this new patent) and also other AI genetic testing.

The lab snaps a standard microscope photo of a day 5 embryo. That stage is called a blastocyst. Think of it as the moment the embryo has a clear outer shell and a growing cell ball inside. 

Software draws a careful outline around the outer shell and the embryo inside. It looks at shape, texture, symmetry, and how expanded the embryo is for its age. Those patterns roll up into a simple score. A higher score means the embryo looks more likely to handle the next steps after transfer. 

That’s "morphology", which is about what you can see in its form and structure. It does not read chromosomes. It does not predict eye color. It cannot tell you about single gene diseases. 

Genetic testing is different. For genetic testing, a small sample is taken from the embryo and sent to a lab. The lab checks chromosome counts and, if needed, looks for known single gene issues. Results help spot embryos that may never lead to a healthy pregnancy or may end early. 

So the split is clear. This patent guides clinicians with speed and zero biopsy. Further genetic testing confirms with direct evidence from the genome. This patent notes that the genetic biopsies can risk causing damage to the embryo, indicating an objective importance of this kind of testing to decrease risk of damage.

In practice, clinics can stack them. Use the photo score to triage a big group fast, then send a smaller set for genetic testing when the family wants that information. It saves time and money while keeping DNA calls where they belong. 

This patent’s AI called embryo viability from a single day-5 photo with about two-thirds accuracy overall. It did better at spotting embryos that would progress than ones that would not, with 74.1% for “viable” calls vs 65.3% for “non-viable.” When they pit the tool against clinicians, the bars show 66.7% for the AI vs 51.0% for clinicians, and in head-to-head disagreements the AI was right 148 times vs 54 for clinicians. 

It is also to remember that a high scoring embryo represents a probability, not a promise. There are many other factors, such as genetic issues, that may contribute to a "high scoring embryo's" success.

The same playbook fits scenarios where a picture hints at near term “fitness.” Think oocyte and sperm quality checks from standard lab images. Think quality control for cell therapies where labs need a quick visual read before a longer assay. Even animal breeding programs already using embryo imaging could get a faster ranker without new machines. 

This technology spreads because it’s cheap, fast, and fits what clinics already do. For patients wary of touching the embryo, that pitch lands.

Regulators are warming up to AI in fertility, but they want proof and guardrails. The FDA says its old playbook wasn’t built for adaptive AI and many changes may need fresh review. n Europe, AI embryo tools are already getting CE Marks, like Alife’s Embryo Predict in October 2025. That’s a green light for use across the EU, but not a free pass in the U.S. (PR Newswire)

Clinics have business reasons, too. PGT-A genetic testing can add roughly $3,000 to $5,000 per cycle in the U.S., and sometimes much more with extra transfers. A fast image score can triage which embryos to send for genetics, lowering cost per decision. (Gaia Family) Some clinics even absorb AI selection as no-charge “table stakes,” betting it speeds decisions and standardizes grading. Others sell imaging add-ons like time-lapse for a fee. Incentives vary by country and clinic.

For this patent’s technology, a strong “look” is a probability signal, not proof. It can miss chromosome problems, and model scores can drift when labs change microscopes or lighting.

Clinics like genetic testing because it measures what this patent can’t. Chromosome errors are a major reason embryos fail or end early, and testing can help avoid some losses. But biopsy adds complexity. Results can be messy when an embryo has mixed cells, called mosaicism. And not every study shows a clear live-birth boost for every patient group. (HFEA)

There are also real ethics questions around genetic testing. Many people worry that DNA testing slides toward picking traits. Even if a clinic only checks chromosomes today, the line can blur when families ask for more. That’s where this patent’s approach gets airtime, because it helps choose a transfer without touching DNA and can act as a first pass before any genetic call is even on the table.

But a “lighter touch” isn’t a free pass. An image model can still bake in bias, overfit to one clinic’s equipment, or overpromise what a score means. If you lean on it alone, you risk skipping embryos that look average but are genetically healthy. (PMC)

First, there are signs of success for technology akin to that in this patent, involving image-based embryo scoring from photos or time-lapse. Fairtility says its CHLOE platform is in more than 100 clinics across nine countries. Presagen’s Life Whisperer is cleared for sale in 47 countries. Alife’s Embryo Predict just earned a CE Mark in October 2025, opening the EU market. Vitrolife’s iDAScore rides on its EmbryoScope time-lapse network, which the company says totals 2,500+ instruments with capacity for over 1.1 million treatment cycles a year. (PMC)

We can contrast this with figures for DNA analysis technology. UK clinic price lists peg PGT-A in the thousands per cycle, with example menus showing minimum charges around £3,500 to £5,500 for analysis plus per-embryo fees, while time-lapse imaging is often a few hundred pounds when billed as an add-on. In the US, publications and patient watchdogs regularly cite PGT-A as a multi-thousand-dollar decision driver per cycle.

Adoption tells its own story. A cross-section of UK clinics advertised time-lapse imaging at scale even before today’s AI overlays, and US registry data show genetic testing has been used in a large share of IVF cycles in recent years.

In 2021, Vitrolife bought Igenomix (a global PGT operator) for €1.25 billion. Igenomix alone reported a network of roughly 19–20 labs serving ~80 countries and assisting about 90,000 couples per year at the time of the sale.

And governments are also contributing funding towards this industry. The EU has funded fertility-adjacent research consortia such as ReproUnion’s 2023–2026 innovation platform with a €6.7 million budget, plus project-level grants like Predict AYA at nearly €7 million and additional fertility research awards in the €1.8–4.0 million range under Horizon Europe. In the US, NIH’s total budget hit roughly $47.4 billion in 2024. (MVA)

When embryo choices get messy, this AI reads a single day-5 photo and gives you a near-term viability score.

That can mean fewer biopsies for some families, quicker triage for labs, and clearer talks in the clinic. It’s still a probability, not a promise.

Where’s your line? Is further DNA testing a must for peace of mind, or does it slide toward picking traits? Who should decide when to test, when to stop, and what “good enough” looks like?

Ready to claim your stake in this shift? Dive into the details: US 2025/0278835 A1.