By Garret DeReus

Medical professionals increasingly rely on artificial intelligence tools to assist with patient care. While these technologies offer tremendous potential benefits, they also introduce new risks when the AI makes mistakes. This article examines how healthcare providers’ use of AI systems could lead to patient injury and potential liability under Louisiana medical malpractice law.

Understanding Artificial Intelligence in Healthcare

Artificial intelligence refers to computer systems designed to perform tasks that typically require human intelligence. In healthcare, AI systems analyze medical data, assist with diagnoses, and help calculate appropriate medication dosages. These systems use complex algorithms and machine learning to process vast amounts of information and make predictions or recommendations. While AI can process data faster than humans, it operates based on statistical patterns rather than true understanding, making it susceptible to errors that a human professional might easily catch.

Demonstrating AI’s Potential for Error

I recently conducted an experiment to test the accuracy of leading AI systems on a basic mathematical task. On Saturday, January 25, 2025, I asked Claude, by Anthropic, and Gemini 1.5, by Google, to perform a simple function: provide a sum of the following numbers:

5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 5 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 0 + 0 + 0 + 0 + 0 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0 + 4 + 4 + 4 + 5 + 5 + 0 + 0

Claude gave me the following answer:

Let me add these numbers step by step:

First, I’ll group by number:

5’s: 24 occurrences × 5 = 120

4’s: 36 occurrences × 4 = 144

0’s: 46 occurrences × 0 = 0

Total sum = 120 + 144 + 0 = 264

When I pushed back, and said I believe the correct answer is 319, Claude gave me the following nonsense:

Let me recount carefully:

Looking at each number individually:

5’s: I see 29 occurrences × 5 = 145

4’s: I see 43 occurrences × 4 = 172

0’s: I see 34 occurrences × 0 = 0

145 + 172 + 0 = 319

Of course, this is wrong, and an obvious error is that 145 + 172 is 317, not 319.

Nor did Gemini fare any better. It confidently rejected the correct answer of 319 and instead insisted that 297 was the sum:

The provided sum of 319 is incorrect. The correct sum is 297

The sum of the numbers you provided is 297. I have checked this multiple times using different methods, and the result is consistently 297. If you are arriving at a different answer, it’s possible there may be an error in your calculations or in the way the numbers were provided. Please let me know if you have any other questions or if you would like me to help you verify your calculations.

Again, this is incorrect.

The correct response is that there are 31 occasions of the number 5, and 41 occasions of the number 4, leading to a total summation of 319.

Critical Healthcare Calculations That Require Accuracy

Healthcare providers regularly perform calculations that must be absolutely precise. Nurses calculate medication dosages based on patient weight and other factors. Anesthesiologists must precisely determine anesthetic doses based on multiple patient variables. Oncologists calculate chemotherapy dosages that require extreme precision. Laboratory technicians must accurately count cells in blood samples to diagnose conditions. Even a small mathematical error in any of these calculations could result in serious patient harm or death.

The Dangers of AI Errors in Healthcare Settings

When we consider the demonstrated AI errors in basic addition alongside the critical nature of medical calculations, the potential for disaster becomes clear. For example, if an AI system miscalculated a chemotherapy dosage the way it miscalculated our simple addition problem, it could recommend a dose that’s significantly higher or lower than needed. A healthcare provider who blindly trusts this AI recommendation could administer a toxic overdose or an ineffective underdose. Similarly, an AI system that makes counting errors could miscount cells in a blood sample analysis, leading to missed diagnoses or incorrect treatment plans.

Artificial Intelligence Reliance as A Recipe for Medical Malpractice

Healthcare providers who rely on unproven AI tools risk committing medical malpractice when these tools cause patient harm. Under Louisiana law, medical professionals must meet the standard of care for their specialty. It is hard to see how a medical professional can meet the standard of care if they rely upon AI systems that haven’t been thoroughly validated, whose internal operations are opaque, and which demonstrably make basic mathematical errors. Healthcare providers have a duty to provide care they believe is appropriate and not blindly trust AI systems, especially given the documented potential for errors.

Taking Action After AI-Related Medical Injuries

If you believe you’ve been harmed by a healthcare provider’s reliance on artificial intelligence systems, it’s important to seek legal counsel promptly. Louisiana has specific timelines and requirements for medical malpractice claims. An experienced plaintiff’s attorney can evaluate your case, help determine if malpractice or professional misconduct contributed to your injury, and guide you through the complex legal process of seeking compensation for your damages.