Insights / AI in Personal Injury Litigation
Why AI Alone Isn’t Enough for Medical Record Review
By Breen LNC.
Short answer: AI alone falls short because reading words is different from weighing them. Studies show AI summaries miss and misstate things, people tend to trust automated output too much, and professional standards keep responsibility with a human. Medical record review needs a clinician who can tell what matters.
The errors are real, and some are quiet
In a study of 100 hospital encounters at UCSF, AI narratives were more concise and coherent than physician narratives, and less comprehensive. They contained more errors per summary, 2.91 against 1.82. The authors concluded that AI narratives with human review may be a viable option. The human review is part of that conclusion.
People over-trust machines
The National Institute of Standards and Technology describes a related risk in its generative AI profile. It calls false output “confabulation,” says it is especially relevant in long responses and in fields that need domain expertise, and warns about automation bias, or “excessive deference to automated systems.” In other words, a confident paragraph is more likely to slide past a tired reader.
The FDA has flagged the same pattern in health care. Its 2026 guidance on clinical decision support software defines automation bias as the tendency to over-rely on a suggestion from an automated system, which can lead to following bad advice or failing to act when not prompted. That guidance governs medical device regulation, not litigation, but the human tendency it describes doesn’t change with the setting.
Judgment has a license behind it
The American Association of Legal Nurse Consultants describes a legal nurse consultant as “a licensed registered nurse who performs a critical analysis” of clinically related issues, applying education and experience to standard of care, causation and damages. Its FAQ lists work such as spotting record tampering, omissions and contradictions.
Those tasks turn on knowing what should be in a chart and isn’t. Software can tell you what a page says. Knowing that a page ought to exist and doesn’t takes clinical experience.
The American Nurses Association’s position on AI points the same way for nursing practice: AI does not replace a nurse’s decision-making, judgment, critical thinking or assessment skills, and nurses remain accountable when they use it.
Use both
Let the software handle volume, and let a nurse decide what the record means.
Frequently asked questions
What is automation bias?
FDA guidance describes it as the tendency to over-rely on a suggestion from an automated system, which can cause people to follow incorrect advice or fail to act when not prompted.
Does AI replace a legal nurse consultant?
The sources reviewed here do not support that. AALNC describes the role as critical clinical analysis by a licensed registered nurse, and the ANA says AI does not replace a nurse’s judgment.
Sources
- Williams et al., Physician- and Large Language Model-Generated Hospital Discharge Summaries, JAMA Internal Medicine (2025)
- NIST AI 600-1, AI Risk Management Framework: Generative AI Profile (July 2024)
- U.S. Food and Drug Administration, Clinical Decision Support Software: Guidance for Industry and FDA Staff (Jan. 29, 2026)
- American Association of Legal Nurse Consultants, LNC Awareness Week
- American Association of Legal Nurse Consultants, FAQs
- American Nurses Association, The Ethical Use of Artificial Intelligence in Nursing Practice, OJIN Vol. 30, No. 2 (May 2025)
General information for legal professionals. This is not legal or medical advice.