close
R I A

Artificial Intelligence in Medicine

Rasa.AI Labs · Rasa Institute of Analytics

A practical 30-hour programme for medical students

From AI literacy and clinical research to medical data, imaging concepts and responsible innovation. Artificial Intelligence in Medicine is for students on MBBS and other recognised medical programmes. It is beginner to intermediate: practical use of AI, with a person still checking the result.

Sessions are instructor-led, with demonstrations, guided labs and a capstone. You use AI as a supervised study aid. You learn to question its output, and you build a small educational prototype on synthetic data, public teaching material, or data the institution has approved and de-identified.

This is education. It is not clinical training, a licence to practise, or authority to diagnose, prescribe, triage or treat a real patient. Classroom batches run at Purasawalkam and Saligramam. Call +91 98843 63200.

Who
Medical students

Duration
30 guided hours

Level
Beginner to intermediate

Delivery
Classes, labs and a capstone

What you will be able to do

  • Explain how artificial intelligence, machine learning, deep learning and generative AI differ.
  • Use an assistant for medical study, a literature pass and an evidence summary, then check the sources.
  • Write a prompt and test the answer against a textbook, a guideline or a PubMed abstract.
  • Turn a simulated consultation into structured notes, and mark what a person still has to correct.
  • Describe what imaging models can and cannot do, including false positives and false negatives.
  • Explore a teaching dataset and make a simple chart or an introductory model.
  • Assemble a retrieval assistant that answers only from approved notes and shows the passage it used.
  • Apply privacy, consent, bias, safety and academic-integrity rules when you use AI.

30 hours

ModuleTopicHours
1AI foundations for medicine3
2Generative AI for medical learning and research4
3Clinical documentation and communication support4
4Medical data, statistics and analytics3
5Medical imaging and computer vision4
6Clinical decision support and precision medicine, as concepts3
7A medical knowledge assistant that retrieves from approved notes4
8Responsible AI, privacy and research integrity2
9Capstone demonstration and assessment3
Total30

A usual plan is ten sessions of three hours, or another timetable the institution approves. A group of about 20 to 30 students suits the labs. Programming helps. It is not required. Selected labs have a path that does not start from code.

Modules

1 — Foundations (3 hours)

Name the kinds of AI and where they appear in study, research, imaging and hospital work. See how a language model helps, where it fails, and where a medical student should not use it. You leave with a one-page map of five fitting uses and five that are out of bounds.

2 — Learning and research (4 hours)

Prompt for anatomy, physiology, pathology and pharmacology at a stated level. Summarise an open-access abstract: design, population, limits. Make flashcards or questions, then verify every answer. Compare claims with PubMed, a guideline or a textbook.

3 — Notes and plain language (4 hours)

On a fictional transcript, draft a SOAP note. Separate what the person said, what was observed, and what the model added. Rewrite a technical point in plain language for a hypothetical reader. Mark omissions, invented detail and unsafe wording.

4 — Data and simple analytics (3 hours)

Work on a synthetic or public teaching table. Counts, proportions, averages, a chart, and what the chart does not prove. Talk through missing values, leakage, sampling bias, confounding and overfitting. Submit three findings you can defend.

5 — Imaging concepts (4 hours)

Image formats, a display, and the ideas of classification, segmentation and detection. A guided notebook on a licensed public educational set. Read a confusion matrix. This module does not ask you to diagnose a real patient.

6 — Decision support, as a concept (3 hours)

Alerts, risk scores and evidence lookup. A fictional case: what a tool might organise, why a score is not a diagnosis, and when a clinician must set the suggestion aside or escalate. You are not authorised to make a care decision from the model.

7 — A knowledge assistant (4 hours)

Chunk approved notes, retrieve a passage, and answer only from that passage. The assistant shows the source and says when the notes do not contain the answer. You test vague questions, unsupported claims and passages that disagree.

8 — Privacy and integrity (2 hours)

What must never be pasted into a public tool. Consent, disclosure and institutional approval. A checklist for AI-assisted academic work: bias, credit, copyright and a person who remains accountable.

Capstone

Three hours: finish one educational prototype, show it, and answer questions. Use synthetic, public, or properly approved de-identified data. State the sources, the limits, the privacy controls, and one change required before any real-world use.

OptionWhat you show
Study companionAnswers from approved lecture notes, with the supporting passage
Paper summariserDesign, population, intervention, outcomes, limits and terms from a paper you were given
Simulated note assistantA fictional transcript turned into a draft note for a person to review
Image learning notebookPreprocessing and an evaluation on an approved educational image set
Dataset explorerCharts and descriptive statistics on a synthetic or public teaching table
Education draft assistantPlain-language material from clinician-approved source text, with citations

How the work is marked

  • Quizzes — 15%
  • Labs — 25%
  • Research and documentation — 15%
  • Capstone prototype — 25%
  • Presentation, viva and responsible-use reasoning — 20%

The capstone itself is marked on the problem (15%), the build (20%), grounding and checks (20%), privacy and safety (20%), clarity (10%), and reflection (15%).

Limits

  • Do not put identifiable patient information into a public AI tool.
  • Synthetic data unless the institution has approved a lawful, safeguarded exception.
  • Check important claims against a textbook, a paper, a guideline or faculty.
  • No model output is a diagnosis, a prescription, a triage or a treatment for a real patient.
  • A live clinical system needs separate governance, validation and approval. This course does not provide that.

Completion: attend at least 80% of the guided hours, submit the required practicals, meet the pass mark RIA sets for the batch, and present the capstone. The certificate wording is an RIA completion record. Call +91 98843 63200.

Enquire about AI in Medicine

Explore related RIA programmes

AI in Healthcare
AI in Biomedical
AI in Pharma
AI in Research
Medical Robotics
Industry hub

Ready to start AI in Medicine?

Enquire for the next 30-hour batch at Purasawalkam or Saligramam. For medical students. Not a diagnosis or prescribing course. Call +91 98843 63200.

ENQUIRE NOW

Across every course

Seminars, workshops and faculty development

The same formats sit on any RIA subject. Name the course and the room — a faculty programme, a seminar, a workshop, an internship, or help with a project.

Faculty programme Seminar Workshop Internship Projects
See how a session is arranged
Go To Top