Future directions in infertility research: the role of generative AI and large language models.

A solo computer scientist argues LLMs could transform male infertility research — with zero data to show for it

Journal: Systems Biology in Reproductive Medicine | Published: 2026-03-16 | Type: Review | PMID: 41838807 Authors: M Jamil Hasan (Department of Computer Science, University of Idaho, Moscow, ID) Funding/COI: Not listed

Summary

A single computational scientist surveys the theoretical potential of large language models and generative AI for male infertility research — genomic interpretation, sperm quality assessment, offspring health prediction via polygenic risk scoring. The paper contains no original data, no experimental results, and no empirical validation of any claim. It reads more like a grant pitch than a scientific review.

Claims

• Up to 50% of infertility cases involve a male factor, with a large proportion still classified as idiopathic
• LLMs could accelerate interpretation of complex genomic data and identify novel gene-environment interactions in male infertility
• Generative AI could enhance sperm quality assessment
• AI could forecast offspring long-term health risks stemming from paternal factors via polygenic risk scoring and in silico developmental simulations
• Interdisciplinary AI approaches could "lower barriers" between computer science and reproductive biology

Study Quality

This is a narrative review with no systematic methodology, no documented search strategy, no PRISMA flow, and no data. The sole author is a computer scientist — the abstract itself concedes it was "written from the perspective of a computational scientist collaborating with fertility experts," but no fertility expert appears anywhere in the author list. That collaboration, if it exists, did not produce a co-author.

The paper offers no evidence that LLMs have actually improved any outcome in male infertility research. Every substantive claim is framed as future potential: "could," "stand poised," "offer a transformative opportunity." The abstract promises "real-world examples and emerging case studies" but none appear in the available data. There are no benchmarks, no comparisons to existing analytical methods, and no reproducible results.

Red Flags

• No funding listed — unusual for a published review; raises questions about institutional support
• No COI disclosed — the author advocates for commercial AI tools without declaring any relationship to AI vendors
• Single computer science author reviewing a biomedical domain; no reproductive medicine co-author
• Zero empirical data of any kind
• Polygenic risk scoring for offspring health via paternal factors is presented without addressing well-documented PRS limitations across diverse populations
• "Transformative opportunity," "stand poised to unlock novel insights" — promotional framing, not scientific language
• Speculative ethics discussion without any concrete ethical framework or risk analysis

Strengths

• Correctly identifies male infertility as understudied relative to female infertility
• The 50% male-factor figure is a defensible estimate consistent with published literature
• Raises legitimate questions about AI interpretability and social responsibility in reproductive medicine
• Interdisciplinary framing has genuine value — when it comes with data

Verdict

This is a speculative opinion piece formatted as a review. The missing funding and COI disclosures are red flags, not minor omissions — an author advocating for AI adoption in medicine should declare any conflicts. The paper's core argument (AI could help) may be true, but arguing it requires evidence, and there is none here. Useful as background reading for someone writing a grant; not useful as a basis for any scientific claim.