AI scans 2.5 million images to help a couple conceive after 19 years and several failed IVF attempts

AI

For 19 years, they tried everything. Multiple IVF cycles, two painful surgeries, and countless failed attempts only to be told there was no sperm to be found. Then came the twist no one expected: a computer found what humans couldn’t.

At Columbia University, researchers have developed an artificial intelligence system that can spot even the rarest sperm cells, those that often go unnoticed, hidden in what looks like a barren sample. The breakthrough, called Sperm Tracking and Recovery (STAR), has already led to its first confirmed pregnancy, and it might just rewrite what’s possible for couples facing severe male infertility.

In this case, the AI scanned nearly 2.5 million microscopic images of the man’s semen sample over two hours and located two viable sperm cells, enough to help the couple conceive. “A semen sample can appear totally normal, but under the microscope, you might see nothing but a sea of debris,” said Dr. Zev Williams, who heads Columbia University’s Fertility Center and led the study.

The story, published in The Lancet, highlights how AI is moving beyond chatbots and automation, it’s now creating new chances at life.

Male infertility contributes to about 40% of all infertility cases, and for men with azoospermia or cryptozoospermia, sperm can be so rare that even skilled embryologists can’t find them. Until now, many of these men were told their only options were donor sperm or adoption.

STAR changes that equation. The AI-powered device scans semen samples at astonishing speed over a million microscopic images every hour, it identifies and isolates potential sperm cells without surgery or invasive procedures. It’s faster, safer, and far more precise than traditional methods.

Researchers believe that if future trials confirm its success, STAR could revolutionize fertility medicine. But for one couple, it already has. After nearly two decades of heartbreak, AI has given them something no machine ever had before — a reason to hope again.