Researchers train AI system to predict which IVF embryos will grow into healthy babies

Scientists have developed a new artificial intelligence system that may accurately predict which IVF embryos will grown into healthy babies. At present embryologists are employed to choose those fertilised eggs that are most likely to result in a pregnancy. This involves checking them by eye to see which appear to be healthy. However, scientists believe the system could be greatly improved by training machines to perform the task. They tested the idea on cow embryos by training the AI to find what a good embryo looks like from a series of images. By removing the potential for human error, more cycles of IVF could result in healthy babies, researchers said. “Instead of a human looking at thousands of images, a piece of software looks at them and is capable of learning all the time,” said Stuart Lavery, consultant at the Boston Place Clinic in London in the UK. “As we get data about which embryos produce a baby, that data will be fed back into the computer and the computer will learn,” Lavery was quoted as saying by The Times. Researchers trained a computer to identify healthy embryos using images of 482 bovine embryos during their development. This enabled the system to spot early flaws that led to poor development - many of which were invisible to humans. Researchers found that the technique was much more consistent than human embryologists. It was more reliable and it could also look for things that the human eye could not see, they said.