healthHealth and Medicine

Artificial Intelligence Better Than Medical Experts At Choosing Viable IVF Embryos


Robin Andrews

Science & Policy Writer

Is AI the future of IVF? Dabarti CGI/Shutterstock

The future of baby-making is set to be very different from the one we have now. Just last week, a researcher boldly claimed that growing embryos in a laboratory setting will become far more commonplace, and will allow us to remove genetic diseases from the equation before the baby is born.

Now, during the annual meeting of the European Society of Human Reproduction and Embryology in Geneva, scientists have given us yet another peek into the future of conception. In a groundbreaking new study, a team of embryologists was pitted against an artificial intelligence (AI) during simulated in vitro fertilization (IVF) selection process – and the AI appeared to be better at selecting viable embryos.


During IVF, an egg is removed from the hopeful mother’s ovaries and fertilized with the potential father’s sperm in a laboratory setting. This fertilized egg is then implanted in the woman’s womb and allowed to develop normally.

It’s used for those with fertility problems, and currently has variable rates of success. Sometimes, the embryos fail for a variety of reasons, and experts are trained to look out for defects that may trigger a failed pregnancy. Between 30 to 60 percent of seemingly viable embryos fail to implant in the uterus.

This new study – a collaborative effort between São Paulo State University and London’s Boston Place Clinic – decided to pit experts against an AI designed to do their jobs for them. Using bovine embryos, the AI was given a chance to train itself to look for viable embryos and highlight defective ones.

Bovine embryos were used this time around, but human equivalents are next. u3d/Shutterstock

Both the AI and a team of embryologists were then given 48 examples of bovine embryos to look at, and had a chance to observe them three times over.


Using just 24 key characteristics, such as morphology, texture, and the quantity and quality of the cells present, the AI was able to pick viable embryos 76 percent of the time. Although the accuracy value for the embryologists was not given, it was said to be lower; importantly, unlike the AI, the embryologists found it difficult getting a consensus on the quality of the embryos.

Overall, the AI was more accurate and consistent in its decision making, with major errors only occurring 6 percent of the time. The team has now moved on to testing the AI out on human embryos to see if its success can be replicated.

According to the team behind the study, physical fatigue, different educational backgrounds, and emotional stress all contribute towards the human experts’ lower accuracy – things that the AI would never be disadvantaged by.

Far from putting the embryologists out of a job, however, the team see this AI as a way to augment their practices.


“The artificial intelligence system must be based on learning from a human being – that is, the experienced embryologists who set the standards of assessment to train the system,” lead investigator Professor José Celso Rocha, from São Paulo State University, said in a statement.

The rise of AI in the last few years has been nothing less than stellar, and already, its ability to trounce humans when it comes to pattern recognition has led to it being used in a wide range of settings, from the military to health services. Certain cities around the world are already trialing the use of an AI to make diagnoses, which is designed to compliment the abilities of a human clinical practitioner.

This new study is yet another confirmation that an AI-led future is far closer to the present than we think: man and machine, working hand in hand, to prolong the human species.

Expect to see more AI involved in the medical world in the near future. Carlos Amarillo/Shutterstock


healthHealth and Medicine
  • tag
  • treatment,

  • ivf,

  • fertility,

  • artificial intelligence,

  • pregnancy,

  • babies,

  • embryos,

  • trial,

  • human experts,

  • bovine,

  • teams