A wellness app deployed an AI chatbot to answer health questions.
A user asked about persistent headaches and dizziness.
The AI analyzed the symptoms and responded: “These symptoms are likely stress-related. Try relaxation techniques, stay hydrated, and get more sleep. You should feel better in a few days.”
The user followed the AI’s advice. Skipped seeing a doctor. The symptoms got worse.
Turned out it was a stroke. Delayed treatment caused permanent damage.
Now the user is suing the company for practicing medicine without a license, negligent medical advice, and personal injury.
The company argues it’s just a wellness app with a disclaimer saying “not medical advice.”
Courts aren’t buying it.
Where Wellness Becomes Medical Advice
Companies think adding “not medical advice” disclaimers protects them.
It doesn’t.
What courts look at: Did the AI provide personalized recommendations based on individual symptoms? Did it tell the person what to do or not do? Did a reasonable person believe they were getting medical guidance? Did the advice relate to diagnosing, treating, or managing health conditions?
If yes to these questions, it’s medical advice regardless of what the disclaimer says.
Wellness tips vs. medical advice: General wellness information (“staying hydrated is healthy”) is fine. Personalized symptom analysis (“your headache is probably nothing serious”) crosses into medical practice.
Most health and wellness apps cross that line without realizing it.
The Practicing Medicine Problem
Every state prohibits practicing medicine without a license.
When AI gives personalized health advice, courts are finding that it constitutes medical practice.
It doesn’t matter that it’s AI: The law prohibits unlicensed medical practice, whether done by humans or algorithms.
The company is liable: The company operating the AI is practicing medicine through its software.
Criminal violations: In some states, unlicensed medical practice is a criminal offense, not just civil liability.
Injury creates damages: When someone gets hurt following AI medical advice, the company faces both regulatory violations and personal injury claims.
Adding a disclaimer doesn’t change whether the AI is practicing medicine.
Real Cases Emerging
The mental health chatbot: An AI therapy app gave personalized mental health advice. The user experienced a crisis. AI recommended coping strategies instead of emergency intervention. User harmed themselves. The family sued for negligent treatment and practicing psychology without a license.
The symptom checker: A health app’s AI analyzed user symptoms and recommended they monitor at home rather than seek immediate care. The user had appendicitis. Delayed treatment led to complications. The user sued for medical negligence.
The medication advice bot: An AI gave advice about medication interactions and dosage adjustments. User followed the advice. Had an adverse reaction requiring hospitalization. User sued for unauthorized pharmaceutical practice.
In each case, companies argued they provided information, not medical advice. Courts found the personalized nature of AI recommendations constituted medical practice.
FDA Regulation Risk
Some AI health tools may qualify as medical devices requiring FDA clearance.
What triggers FDA regulation: AI that diagnoses conditions, recommends treatment, or analyzes patient data to inform clinical decisions may be a medical device.
What companies miss: They think their app is wellness software. The FDA may classify it as an uncleared medical device.
Enforcement risk: Operating an uncleared medical device violates federal law and creates additional liability beyond state medical practice violations.
Most wellness apps have never conducted an FDA analysis to determine if their AI requires clearance.
The Disclaimer Problem
Every health app has disclaimers saying “consult your doctor” and “not medical advice.”
These disclaimers don’t eliminate liability when the AI actually provides medical advice.
Courts ignore disclaimers when: The AI functionality contradicts the disclaimer (saying “not medical advice” while giving personalized treatment recommendations), users reasonably rely on AI recommendations as medical guidance, and the disclaimer is buried in terms of service users don’t read.
What disclaimers can’t do: Authorize unlicensed medical practice, eliminate the duty of care when AI gives dangerous advice, or prevent personal injury claims when users get hurt.
Disclaimers create a false sense of legal protection.
What Health and Wellness Companies Should Do
If companies offer AI that interacts with users about health, they need to understand the legal boundaries.
Get legal review: Have healthcare lawyers review what the AI does and whether it crosses into medical practice.
Limit AI scope: Configure AI to provide only general wellness information, not personalized medical recommendations.
Implement safeguards: Program AI to always recommend consulting a doctor for specific symptoms, refuse to diagnose or recommend treatment, and escalate concerning symptoms to human medical professionals.
Consider FDA compliance: Determine if the AI qualifies as a medical device requiring clearance.
Involve licensed professionals: If providing personalized health guidance, ensure licensed medical professionals supervise the AI.
Document limitations: Clearly show users what AI can and can’t do, and when they need real medical care.
Most health apps did none of this because they assumed wellness apps don’t face medical liability.
Ask Yourself This
Does your company offer AI that answers health questions, analyzes symptoms, or provides personalized wellness recommendations?
If yes, has anyone determined whether that constitutes medical practice requiring licensure?
What happens if someone follows AI health advice and gets hurt?
Most companies can’t answer because they never considered that their wellness app might be practicing medicine.
Schedule a 30-minute AI Health App Risk Assessment where we’ll:
- Review whether your AI crosses into medical practice
- Show what medical licensure and FDA requirements may apply
- Identify liability gaps in current AI health features
- Help implement safeguards to limit medical practice risk
- Create protocols for when AI encounters serious symptoms
Because offering AI health advice feels innovative until someone gets hurt following the recommendations, and you discover wellness apps can practice medicine illegally.
Brian Alexander
Founding Partner, VisionAI+ Consulting Group
Former General Counsel, AI Engineering Firm
Certified ISO 42001 Lead Auditor