Our AI Midwife

A couple struggling with six years of unexplained infertility describes turning to a large language model as a constant, on-demand medical adviser. By coaxing the model into role-playing a TV diagnostician named Dr. Reid, they asked detailed questions about hormone levels, scans, and surgical options; the AI’s answers generally matched their doctors’ guidance, though it sometimes hallucinated or resisted certain prompts until rewarded or reformulated. The AI suggested an MRI that clinicians had not emphasized; that scan revealed a 6 cm submucosal fibroid missed for years, which was surgically removed and followed almost immediately by a natural pregnancy.

The piece argues that consumer AI can meaningfully augment patient agency by filling time- and expertise-related gaps in clinical encounters: faster, patient-driven hypothesis generation, targeted testing, and curated specialist referrals. It acknowledges real risks - misinformation, model safety limits, and social backlash - while insisting those concerns don’t erase concrete benefits in personal medical outcomes. The author presents AI as a controversial but powerful “midwife” that helped identify and solve a hidden problem, closing with a personal note that their son was named Reed after the model’s fictional persona.