The Role of AI Chatbots and Large Language Models in Patient Engagement, Support, and Healthcare Operations
Executive Summary
Artificial intelligence (AI) chatbots—especially large language model (LLM)-based systems such as ChatGPT—are rapidly transforming patient engagement across healthcare. Literature published between 2023 and 2025 shows accelerated adoption for symptom checking, health information seeking, appointment scheduling, administrative assistance, and communication support. This whitepaper synthesizes findings from peer‑reviewed studies, systematic reviews, and healthcare reports to provide a cohesive understanding of how AI chatbots are currently used, their benefits, limitations, and strategic implications for healthcare organizations.
1. Introduction
Digital transformation in healthcare is shifting patient expectations toward immediate, reliable, and personalized interaction. AI chatbots and LLMs now serve as patient-facing systems that provide symptom triage, answer FAQs, simplify medical information, and facilitate administrative tasks. As adoption increases, health systems seek evidence-based guidance to deploy these tools safely, effectively, and ethically.
2. Overview of AI Chatbots in Healthcare
Research indicates that chatbots now support four major domains of patient interaction:
Symptom checking & triage – guiding patients through self‑assessment and directing them to appropriate care services.
Patient education & FAQs – providing accessible explanations, medical clarifications, and health-condition insights.
Administrative support – managing appointments, reminders, prescription refills, and portal navigation.
Communication augmentation – simplifying radiology reports, preparing patients for visits, and reducing friction during sensitive encounters.
3. Evidence Review from Key Articles
3.1 Patient Engagement Patterns
Studies show rising patient willingness to use AI chatbots for core healthcare interactions. Engagement levels are influenced by perceived accuracy, convenience, empathy, and privacy.
3.2 Symptom Checking and Self‑Diagnosis
Findings show strong patient interest in LLM-based symptom-checking tools. However, safety concerns underscore the importance of oversight, clear disclaimers, and escalation workflows.
3.3 Health Information Seeking
Demographic variations are pronounced: younger adults heavily rely on chatbots for quick health information, while older adults prefer hybrid models combining digital tools with human support.
3.4 Administrative Assistance
Chatbots demonstrate the highest adoption in administrative tasks such as scheduling, refills, and reminders. Observational studies from health systems show significant efficiency gains.
3.5 Clinical Communication Support
LLMs have proven effective in simplifying radiology reports and enhancing patient comprehension, reducing communication barriers.
4. Key Findings and Trends
4. 1. Symptom checking / self-diagnosis
In a 2023 survey of 607 people, around 78% said they would use ChatGPT to self-diagnose health issues.
There is very high stated willingness to use ChatGPT-style tools for symptom checking/self-triage, even though clinical accuracy and safety remain concerns. This supports building symptom-checker flows (with strong disclaimers and escalation to clinicians).
4.2. Ongoing health information & FAQs
A 2024 KFF poll found 17% of adults use AI chatbots at least once a month to get health information and advice, rising to 25% among 18–29-year-olds.
Roughly 1 in 6 adults are already using AI chatbots as a regular health FAQ channel, with much higher usage in younger cohorts. Any patient portal aimed at under-40s can assume a sizable chunk of users are “AI-native” already.
4.3. Comfort using chatbots for appointment booking (sensitive issues)
A 2024 healthcare consumer survey reported 66% of US patients with sensitive health issues are more comfortable scheduling appointments via an online chatbot than with human staff.
Chatbots reduce embarrassment / stigma friction in access. For sexual health, mental health, and other sensitive services, chatbot-first appointment flows could meaningfully increase uptake and earlier presentation.
4.4. Willingness to use healthcare chatbots for basic admin (appointments, refills, reminders)
A 2025 Accenture-reported survey found 77% of patients would use a healthcare chatbot for basic tasks such as making appointments, prescription refills, and visit reminders.
Most patients are open to self-service, chatbot-based admin as long as it’s simple and reliable. That’s strong justification for using ChatGPT-style flows for FAQs and appointment logistics (slots, prep instructions, reminders).
4.5. Real-world usage of AI symptom/triage tools
At Cedars-Sinai (US), the Cedars-Sinai Connect AI platform (chatbot-based intake, symptom assessment and treatment suggestions) has been used by over 42,000 patients since its 2023 launch.
Large healthcare systems are already running AI-enabled triage at tens of thousands of patients scale. It’s feasible to integrate LLM/chatbot flows (including ChatGPT-powered systems) into mainstream patient engagement safely, with clinician oversight.
4.6. What patients actually use chatbots for today
A 2023 survey (reported via BusinessWire) of doctors and healthcare leaders noted that, among current patient chatbot use cases, the top three were:
72%: scheduling appointments
66%: requesting prescription refills
63%: accessing stored medical data (e.g., history)
In live deployments, appointment scheduling is already the #1 chatbot use-case. That aligns strongly with using ChatGPT-style flows for availability queries, booking, rescheduling, and visit prep FAQs.
7. Cedars-Sinai's AI tool delivered 24/7 care to 42,000 patients. Now, doctors can focus more on treatment, less on paperwork.
5. Benefits of AI Chatbots in Patient Engagement
5.1 Increased Access and Reduced Friction
Patients can obtain immediate answers without wait times.
5.2 Reduced Administrative Burden on Staff
Scheduling, reminders, and routine questions can be automated, freeing staff for higher-value care.
5.3 Enhanced Health Literacy
LLMs excel at translating complex medical information into accessible language.
5.4 Improved Patient Satisfaction
Studies show patients appreciate the anonymity and convenience offered by chatbots, particularly for sensitive conditions.
6. Risks and Limitations
6.1 Potential for Incorrect or Unsafe Advice
LLMs may generate inaccurate medical recommendations, necessitating strong oversight.
6.2 Privacy and Security Concerns
Patients may hesitate to disclose sensitive symptoms without assurances of data security.
6.3 Over-reliance on AI
Patients may treat chatbots as substitutes for clinicians; guardrails must mitigate this risk.
6.4 Equity Challenges
Not all patient groups have equal digital literacy or access.
7. Implementation Recommendations for Health Systems
7.1 Adopt a Hybrid Clinical Oversight Model
Ensure clinicians can review, correct, and supervise chatbot outputs for safety.
7.2 Provide Clear Disclaimers and Escalation Paths
Every patient-facing interaction should indicate when a handoff to a human clinician is recommended.
7.3 Integrate with Existing EHR and Portal Workflows
Avoid standalone deployments; integrate scheduling, refills, and reminders into existing systems.
7.4 Monitor Performance with Real-Time Analytics
Track accuracy, completion rates, escalation patterns, and patient satisfaction.
7.5 Conduct Regular Bias, Safety, and Quality Audits
Evaluate chatbot performance across diverse patient groups and conditions.
8. Future Directions
8.1 Personalized Engagement
LLMs will enable personalized care paths using patient history and preferences.
8.2 Multimodal Patient Interactions
Future chatbots will combine voice, text, images, and medical record extraction.
8.3 Regulatory Evolution
Expect tightening regulatory frameworks governing AI in healthcare, including FDA oversight and data governance standards.
9. Conclusion
AI chatbots and LLMs like ChatGPT are reshaping how patients access care, information, and support. Evidence from recent literature demonstrates clear benefits for engagement, satisfaction, and operational efficiency. However, safe deployment depends on transparency, clinical oversight, and ethical design. Healthcare organizations adopting these tools must balance innovation with patient safety and trust.