The Role of AI in Fertility Treatment: Insights from Experts
As technology becomes intertwined with healthcare, many patients are turning to AI tools for assistance with sensitive issues like fertility treatment. However, experts warn that while AI can provide information, it lacks the nuanced understanding that seasoned physicians acquire through years of experience.
Understanding the Limitations of AI in Medical Decisions
Dr. Jaime Knopman, a fertility specialist in midtown Manhattan, emphasizes that AI-generated information might not always be tailored to individual cases. For example, when couples consider in vitro fertilization (IVF), they often receive viability scores for their embryos. While querying AI models like OpenAI’s ChatGPT for next steps based on these scores could yield helpful answers, it misses significant contextual elements. Factors such as the timing of the embryo biopsy, the patient’s uterine lining condition, and previous treatment outcomes are all crucial to informed decision-making.
Moreover, patients sometimes arrive at clinics insisting on a specific course of action recommended by AI. Dr. Knopman notes that while the approach suggested may be common, it may not be the best fit for every individual. “The science is what we study, but then there’s the art of medicine,” she explains. Making decisions in healthcare often requires a deeper understanding than what AI can offer.
The Evolution of AI in Healthcare
Recognizing the gaps in AI-generated medical advice, companies like OpenAI are taking steps to enhance the reliability of their systems. In May 2025, OpenAI introduced HealthBench, a tool developed with contributions from over 260 physicians across 60 countries. This system evaluates AI responses to health-related inquiries by simulating conversations and providing scoring metrics designed by healthcare professionals.
Even with advancements, OpenAI admits that while recent iterations of their models, such as GPT-4.1, can match or lead human doctors in specific tasks, they still struggle with contextual understanding. The challenge remains for AI to provide reliable, personalized insights consistently.
In contrast, Microsoft has launched the MAI Diagnostic Orchestrator (MAI-DxO), reportedly outperforming human doctors in diagnostic accuracy by leveraging multiple AI models, including those from Google and Meta. This innovative approach attempts to combine strengths from various large language models (LLMs) into one cohesive diagnostic tool, emulating the collaboration of human specialists.
As these advanced AI tools become more prevalent, medical education is adapting. Institutions like Harvard Medical School are now integrating AI training into their curriculum, equipping future doctors to leverage technology effectively. Dean Bernard S. Chang reflects on the parallel with the rise of internet usage for obtaining medical information over the past two decades. Just as the internet became an indispensable resource, AI technology is now seen as a vital tool in modern medical practice.
Ultimately, while AI continues to evolve and demonstrates promising advancements in healthcare, the irreplaceable value of human experience in medical decision-making remains crucial. Physicians like Dr. Knopman exemplify the blend of scientific training and compassionate care that technology alone cannot replicate. As both AI and human practitioners learn to coexist, a balanced approach may lead to improved patient outcomes in the complex world of fertility treatment.