Our AI Midwife

Drew Housman recounts a six-year infertility struggle and how a large language model became an indispensable diagnostic companion. Frustrated by brief clinic visits and a diagnosis of “unexplained infertility,” he and his wife began interrogating GPT-4 (framed as a fictional Dr. Reid) about scans, hormones, follicle counts, inflammation, and surgical options. The AI often matched human doctors’ answers, was patient and endlessly available, and sometimes stumbled or ran afoul of platform rules, but it consistently prompted deeper investigation when clinicians seemed ready to move on.

The AI suggested ordering an MRI that clinicians had not emphasized; the scan revealed a 6 cm submucosal fibroid at a common implantation site that prior testing missed. After laparoscopic myomectomy the wife conceived naturally. Housman uses the episode to argue that, despite legitimate safety and ethical concerns about AI, practical benefits already matter to ordinary people - AI acted as an unlikely “midwife” that changed their outcome - while also acknowledging ambivalence and the need for safer development. He closes with a personal note that the couple named their son Reed, a nod to the AI doctor that helped them.