Prerequisite: A course on Artificial Intelligence or Machine Learning or Deep Learning


  1. Artificial Intelligence Fundamentals. (3 hours)
  2. Introduction to responsible AI. (3 hours)
    • Need for ethics in AI. AI for Society and Humanity
  3. Fairness and Bias (9 hours)
    • Sources of Biases
    • Exploratory data analysis, limitation of a dataset
    • Preprocessing, inprocessing and postprocessing to remove bias
    • Group fairness and Individual fairness
    • Counterfactual fairness
  4. Interpretability and explainability (9 hours)
    • Interpretability through simplification and visualization
    • Intrinsic interpretable methods
    • Post Hoc interpretability
    • Explainability through causality
    • Model agnostic Interpretation
  5. Ethics and Accountability (3 hours)
    • Auditing AI models, fairness assessment
    • Principles for ethical practices
  6. Privacy preservation (9 hours)
    • Attack models
    • Privacy-preserving Learning
    • Differential privacy
    • Federated learning
  7. Case study (Any three) (6 hours)
    • Recommendation systems
    • Medical diagnosis
    • Hiring/ Education
    • Computer Vision
    • Natural Language Processing

Course Objectives

The objective of the course is to know about the responsibility of artificial intelligence (AI) to make AI more useful for society and humanity. The course will also teach principles and practices to perform responsible AI.

Learning Outcomes

  • To be able to state aspects of responsible AI such as fairness, accountability, bias, privacy etc.
  • To be able to assess the fairness and ethics of AI modules.
  • To be able to enforce fairness in models and remove bias in data.
  • To be able to preserve the privacy of individuals while learning from them.
  • To be able to develop responsible AI modules for given practical problems and estimate the tradeoff with accuracy.

Text Books

  1. Virginia Dignum, “Responsible Artificial Intelligence: How to Develop and Use AI in a Responsible Way” Springer Nature, 04-Nov-2019;ISBN-10 : 3030303705, ISBN-13 : 978-3030303709
  2. Christoph Molnar “Interpretable Machine Learning”.Lulu, 1st edition, March 24, 2019; eBook. ISBN-10 : 0244768528, ISBN-13 : 978-0244768522 [available online]


The instructor will share the research paper as a reference when required.

Past Offerings

(Note: Past offerings could be under a different course number.)
  • Offered in Jan-May, 2022 by Sahely Bhadra