← Back

What Are Virtual Patients, and What Does AI Change About Them for Learning and Assessment?

Published · By Making Synapses

AI in Medicine & Education

Virtual patients are computer simulations of real clinical cases in which you play the clinician. Classic versions follow a script or a menu of choices, while the virtual patients AI now powers answer free-form questions by voice or chat and react to what you actually say. Used with clear objectives and feedback, they support both learning and assessment.

What are virtual patients?

A virtual patient is a simulated patient you interview, examine and manage on a screen or in a headset, in the role of the health care provider. The Society for Simulation in Healthcare (SSH) gives the broadest meaning in its Healthcare Simulation Dictionary, published with the Agency for Healthcare Research and Quality (AHRQ): "A representation of an actual patient. Virtual patients can take many forms such as software-based physiological simulators, simulated patients, physical manikins, and simulators."

That breadth is why the term can be confusing. A review in BMC Medical Education found that the term "virtual patient" has been used for many years, "but its meaning varies, leading to confusion." The authors classified 536 articles and found that 330 used it in education. Of those, the most common form was the Interactive Patient Scenario (37%), followed by High Fidelity Software Simulations (19%) and Virtual Standardized Patients (16%).

In everyday use, students and educators usually mean two things:

  1. Interactive patient scenarios. A case on screen where you choose questions, tests and decisions, step by step.
  2. Virtual standardized patients. A character you talk to, who answers like a real patient would, so you can practice the interview itself.

AI matters most for the second kind. For a full introduction to the format and its evidence, see what virtual patient simulation is.

What do AI virtual patients change compared with classic virtual patients?

AI virtual patients let you ask anything in your own words and get a fitting answer, instead of picking from a list of prewritten questions. That single change turns a branching case into a real conversation.

Classic virtual patients were built as branches: the author wrote every question and every answer in advance. A 2024 systematic review in Healthcare described those designs as "semi-conversational environments (typically involving option selection from lists rather than using natural language)," and excluded them because they offer less authenticity. The same review notes that "More recently, the advent of AI has facilitated conversational simulations, allowing students to interact with patients in real-time," and that student satisfaction tends to be higher when interactions are more realistic, "often due to the use of artificial intelligence (AI) and natural language processing (NLP) in the simulators."

A Mayo Clinic study in the Journal of Medical Internet Research puts the contrast plainly. After physicians rated 60 conversations with AI virtual patients, the authors wrote: "The unscripted, responsive dialogues contrast sharply with most existing VPs, for which authentic and flexible dialogue is notoriously difficult to replicate and often not attempted." The feature raters most often credited with authenticity was the patient expressing a preference or making a clear choice, noted in 106 of 180 ratings (58.9%).

Here is what changes in practice:

  1. Free-form questions. You phrase the question yourself, so the case tests whether you know what to ask, not whether you can recognize it on a menu.
  2. Dynamic answers. Dynamic AI virtual patients answer follow-up questions, clarify, push back and express worries, much like a person in clinic.
  3. Natural order. You decide the sequence of the interview, which is exactly what an examiner watches.
  4. Richer feedback. A University of Tübingen team in JMIR Medical Education notes that "traditional virtual patients often yield low levels of feedback," while AI can comment on the specific topics you covered or missed.

Can you talk to an AI virtual patient by voice or chat?

Yes. AI virtual patients run as voice or chat encounters, and some tools offer both. The Healthcare review only included patients that were "conversational, either by voice or by keyboard, using natural language in all cases."

Chat is still the most common format in research. A 2025 scoping review of 28 studies in the Journal of Medical Internet Research found that the studies used natural language "including text (n=14) and voice (n=7), with 1 study supporting both modes." Text works well: the Mayo Clinic authors observed that their patients "used only written text; however, authenticity was high even with this limitation."

Each mode trains something slightly different:

  1. Typing (chat encounters). Good for quiet study and for reviewing your exact wording afterward.
  2. Speaking (voice encounters). Closer to an OSCE station: you hear your own pacing and phrasing, and you cannot edit a question once it is out.

A practical routine is to learn a new case type in chat, then repeat it out loud until the sequence feels natural.

What questions should you ask a virtual patient?

Ask the same questions you would ask a real patient, in a logical order and one at a time. An AI virtual patient rewards a structured history, and it exposes gaps fast.

The Association of American Medical Colleges sets the target in its Core EPA guide: "Learners need to integrate the scientific foundations of medicine with clinical reasoning skills to guide their information gathering." It describes the early learner as showing "errors of omission or commission in gathering information." In its example of a learner who is not yet ready, the student takes a full history but "does not include occupational or travel history."

Use this question checklist with any virtual patient:

  1. Opening. Introduce yourself, confirm the patient's name and date of birth, and ask permission to begin.
  2. Chief complaint. "What brings you in today?" Then let the patient talk before you narrow down.
  3. History of present illness. Onset, location, duration, character, what makes it better or worse, radiation, timing and severity.
  4. Associated symptoms and red flags. Ask about the symptoms that would change your plan for this complaint.
  5. Past medical and surgical history. Prior episodes, chronic conditions, hospital stays and operations.
  6. Medications and allergies. Include over-the-counter drugs and supplements, and the reaction for each allergy.
  7. Family history. Conditions in parents and siblings that matter for this complaint.
  8. Social history. Smoking, alcohol, drugs, living situation, occupation and recent travel.
  9. The patient's perspective. "What do you think is causing this?" and "What worries you most?"
  10. Review of systems. A short screen of the other systems.
  11. Summary and check. Summarize what you heard and ask whether you missed anything.

How you ask matters as much as what you ask. In the Tübingen study, students asked a median of 18 questions per conversation, most of them short and clear. Still, 6.3% of questions used medical jargon, 3.3% were chain questions that asked more than one thing at once, and 1.2% were single words such as "smoking." Across conversations, the introduction was covered in 69.6% of interviews, the main complaint in 52.7% and a system assessment in only 29.7%. A University of Münster trial in BMC Medical Education also found that some students "learned" to combine several questions into one message, and the authors note that "asking multiple questions in a row does not reflect a realistic scenario during medical history conversations."

The Münster team scored interviews with an inventory that makes a useful self-check. Its items include "Taking the lead in the conversation," "Specifying symptoms," "Putting questions in a logical order," "Checking with the patient" and "Summarizing." If your interview hits all five, you are practicing the routine examiners look for.

How are virtual patients used for learning and assessment?

Virtual patients are used for learning when the goal is practice with feedback, and for assessment when performance is scored against preset criteria. AI makes both more conversational.

For learning, the point is repetition with reflection. The National Council of State Boards of Nursing calls simulation "A technique, not a technology" in its Simulation Guidelines for Prelicensure Nursing Education Programs, and asks faculty to create a learning environment that "encourages active learning, repetitive practice, and reflection." The INACSL Healthcare Simulation Standards of Best Practice add that "Simulation-based experiences are purposefully designed to meet identified objectives and optimize achievement of expected outcomes."

AI virtual patients fit that model well. In the Münster trial, 21 medical students each worked through four AI patient cases, with six minutes per history. Students rated the simulations as accurate (8.67 out of 10), feasible (8.38 out of 10) and realistic (7.81 out of 10). The scoping review adds that "The system enables anyone to train at any time, unrestricted by time or spatial limitations."

For assessment, the SSH dictionary separates two kinds:

  1. Formative assessment, "a process for an individual or group engaged in a simulation activity to provide constructive feedback for that individual or group to improve."
  2. Summative assessment, "a process for determining the competence of a participant engaged in health care activity."

INACSL confirms that "Simulation-based experiences may include evaluation of the learner." With a virtual patient, every learner meets the same case and the same checklist, which makes practice exams fair and repeatable. The scoping review notes that "highly interactive virtual patients allow educators to better assess students' clinical reasoning skills by directly observing their abilities." A good AI case uses a fixed, step-by-step checklist, so the conversation can feel natural while the scoring stays consistent.

A simple way to combine both uses:

  1. Teach. Run a new case with hints and read every line of feedback.
  2. Train. Repeat it without hints until your score is stable.
  3. Test. Take a random case under time pressure.

Can nursing and EMS students use virtual simulation with AI patients online for free?

Yes. Virtual simulation with AI patients works for any program that trains history taking and patient assessment, and several tools offer free cases to start.

Nursing programs already use simulation as a core teaching method, and the NCSBN guidelines focus on the quality of the experience. An AI patient case with clear objectives and feedback is a strong way to rehearse a focused assessment before clinicals. See the nursing program page for how this connects to NCLEX preparation.

EMS students can drill the patient interview and scene communication before skills testing (see the EMS program page), and PA students rehearse the same history and physical routine as medical students (see the PA program page). For more free options, see where to find an AI patient simulator online free.

How does Making Synapses use AI virtual patients?

Making Synapses, the medical simulator, runs every case with an AI virtual patient. You talk to the AI patient, by voice or by typing, and the AI patient answers. Everything in Making Synapses is AI: the patient, the checklist correction, the score and the feedback from the Making Synapses AI Attending.

Making Synapses is free to start: you get 10 free cases to play, and each case you play uses one. After that, 3 cases stay free forever. A subscription unlocks every case, and a subscriber can share their cases with friends or classmates in multiplayer, as long as the subscriber hosts the session and stays connected.

Each case is an OSCE station: one AI patient, a step-by-step checklist corrected by AI, a score from 0 to 100% and instant feedback from the Making Synapses AI Attending. You examine with your hands and with instruments such as a stethoscope, reflex hammer and penlight, on PC with mouse and keyboard or in VR with a laser pointer. Teach, Train and Test modes follow the learning and assessment cycle above, and the OSCE exam button picks a random patient from your program in Test mode, with no hints and a time limit.

Making Synapses has 726 AI clinical cases in 5 languages for Medicine (MD, DO), Physician Assistant (PA), Nursing (RN, PN, NP) and EMS (EMT, Paramedic). Cases are built around the USMLE Content Outline, COMLEX-USA Master Blueprint, AAMC Core EPAs, NCCPA PANCE Blueprint, NCLEX-RN and PN 2026 Test Plans, National EMS Education Standards and NREMT, and the INACSL Standards, and each case shows the exam areas it trains. Every attempt is saved to your account. Educators edit cases on the web, the change reaches the simulator live, and they see results per student, so every new generation of clinicians can rehearse the encounter before it counts. More details are on the for students page, the OSCE exam page and the Making Synapses FAQ.

FAQ

Are there free AI virtual patients?

Yes. Making Synapses is free to start with 10 free AI patient cases, no card required. Once you use them, 3 cases stay free forever. You talk to the AI patient by voice or typing, examine, and get an AI-scored checklist with instant feedback. A subscription unlocks every case, and a subscriber can share their cases with friends or classmates in multiplayer, as long as the subscriber hosts the session and stays connected.

Is there a medical AI like ChatGPT?

Yes. General AI chat assistants can discuss medical topics, and researchers use the same kind of technology to build AI virtual patients. A 2025 scoping review found that 95.7% of the studies it analyzed used OpenAI's GPT models. A virtual patient is different from an assistant, though: it plays a patient with a defined case, and you are the clinician.

What does virtual patient mean?

A virtual patient is a representation of an actual patient, usually software you interview and examine in the role of the clinician. The term covers simple interactive case scenarios, conversational characters and full simulations, so it helps to ask which kind a tool offers.

Is a patient simulator game the same as a virtual patient?

Often, yes. Many patient simulator games are virtual patients with scores and replays. What matters for learning is whether the game follows real clinical steps and gives specific feedback.

What should you ask an AI virtual patient first?

Introduce yourself, confirm the patient's identity, then ask an open question such as "What brings you in today?" Let the patient answer fully before you move to focused questions.

Can AI virtual patients be used for assessment?

Yes, especially for formative assessment and practice exams. A virtual patient gives every learner the same case and scores the same checklist steps, and timed modes without hints recreate the conditions of an OSCE station.

Sources

  1. Society for Simulation in Healthcare; Lioce L, ed. Healthcare Simulation Dictionary, Third Edition. Agency for Healthcare Research and Quality, AHRQ Publication No. 24-0077, 2025.
  2. Association of American Medical Colleges. Core Entrustable Professional Activities for Entering Residency: Faculty and Learners' Guide. AAMC, 2014.
  3. National Council of State Boards of Nursing. NCSBN Simulation Guidelines for Prelicensure Nursing Education Programs. NCSBN, 2016.
  4. International Nursing Association for Clinical Simulation and Learning. Healthcare Simulation Standards of Best Practice. INACSL, 2025.
  5. Kononowicz AA, Zary N, Edelbring S, Corral J, Hege I. Virtual patients - what are we talking about? A framework to classify the meanings of the term in healthcare education. BMC Medical Education, 2015.
  6. García-Torres D, Vicente Ripoll MA, Fernández Peris C, Mira Solves JJ. Enhancing Clinical Reasoning with Virtual Patients: A Hybrid Systematic Review Combining Human Reviewers and ChatGPT. Healthcare, 2024.
  7. Cook DA, Overgaard J, Pankratz VS, et al. Virtual Patients Using Large Language Models: Scalable, Contextualized Simulation of Clinician-Patient Dialogue With Feedback. Journal of Medical Internet Research, 2025.
  8. Zeng J, Qi W, Shen S, et al. Embracing the Future of Medical Education With Large Language Model-Based Virtual Patients: Scoping Review. Journal of Medical Internet Research, 2025.
  9. Holderried F, Stegemann-Philipps C, Herrmann-Werner A, et al. A Language Model-Powered Simulated Patient With Automated Feedback for History Taking: Prospective Study. JMIR Medical Education, 2024.
  10. Brügge E, Ricchizzi S, Arenbeck M, et al. Large language models improve clinical decision making of medical students through patient simulation and structured feedback: a randomized controlled trial. BMC Medical Education, 2024.