The Double-Edged Sword of AI Double-Edged Sword
by Monalisa Tailor, MD
How We’ re Using AI: Promise, Pitfalls & Practical Use
Artificial intelligence has been both fascinating and unsettling to learn about. It holds tremendous promise, yet it also carries real risk, particularly the risk of confidently presenting incorrect or incomplete information.
One study examining endoscopists who used AI assistance for three months found that when the tool was withdrawn, their ability to independently characterize polyps declined. That finding highlights an important concern: as we integrate AI into practice, we must remain vigilant about preserving our own clinical judgment and diagnostic skills.
As an internist, I’ ve found AI particularly helpful during acute visits. It can streamline documentation and organize information efficiently. However, careful review is essential. AI may“ hallucinate” details the patient never stated, or omit elements that I consider clinically meaningful, like someone who reported being kicked out of the bedroom because the patient’ s nighttime coughing was so disruptive. Subtle contextual details matter in medicine, and AI does not always recognize their importance. Properly guiding and editing AI-generated content has its own learning curve.
Where AI has been especially helpful is in generating patient instructions, particularly in other languages. Clear communication improves outcomes, and AI has made that process more accessible and efficient. I’ ve also seen creative uses: a resident I worked with encouraged patients to use AI tools to generate healthier recipes. It’ s a practical and empowering idea.
In administrative work, AI has been invaluable. Drafting insurance appeals or advocacy letters and then refining them with AI has helped strengthen arguments and add clarity and detail. Rather than replacing my voice, it enhances it.
For clinical research and point-of-care learning, I’ ve used tools like DoximityGPT and OpenEvidence to ask focused clinical questions. The ability to receive concise summaries with citations to peer-reviewed literature has been particularly useful. Just last week, I used AI to update myself on treatment approaches for restless legs syndrome and was directed to a recent JAMA article from January 2026 outlining current recommendations. Having rapid access to evolving evidence supports better clinical decision-making.
Outside of work, AI has also made its way into everyday life. Google’ s integration of AI summaries into search results is sometimes genuinely helpful. I’ ve found it especially useful for travel planning, creating itineraries that help ensure I don’ t miss key experiences.
AI is neither wholly revolutionary nor wholly dangerous. It is a tool: powerful, imperfect and evolving. Like any tool in medicine, its value depends on how thoughtfully we use it. The responsibility remains ours: to question, to verify and to apply judgment.
( I had this article enhanced by AI before Dr. Barry provided her revisions.
Her comment was,“ What’ s AI’ s beef with commas? Commas are useful.”)
Dr. Tailor is a practicing internal medicine physician at Norton Community Medical Associates: Barret.
14 LOUISVILLE MEDICINE