Efter- og videreuddannelse
Kursussøgning

AI for Humanity: Machine Decisions, Learning and Societal Consequences

Course content

This course provides students with a comprehensive foundation in state-of-the-art machine learning methods, their practical and methodological applications, and their societal consequences. The course focuses equally on technical expertise and critical analysis of machine decision-making in various societal contexts, ensuring interdisciplinary appeal to computational and social science students alike.

 

The curriculum is structured into three interconnected parts:

 

Part I: Foundations of Machine Learning

  • Core Techniques: Overview of tree-based models, neural networks, and advanced architectures like transformers.
  • Recent Advances: Introduction to generative AI (e.g., GPT models) and self-supervised learning.

 

Part II: Machine Learning for Policy Evaluation and Econometrics

  • Causal Inference Techniques: Hybrid econometric models leveraging machine learning, such as double machine learning and causal forests.
  • Emerging Methods: Deep instrumental variables (Deep IV), adversarial methods for causal inference, and high-dimensional data analysis.

 

Part III: Algorithmic Policies and Consequences of Adoption

  • Policy and Prediction Frameworks: Introduction to the prediction policy problem and how algorithms shape decision-making processes.
  • Fairness and Bias: Analysis of fairness metrics, unintended consequences, and societal trade-offs in algorithmic outcomes.
  • Challenges in Algorithmic Governance: Exploring transparency, accountability, explainability, and ethical implications of AI deployment in sectors like healthcare, finance, and public policy.
  • Societal consequences of AI adoption: labor market dynamics, income inequality and educational learning

 

Throughout the course there is an emphasis on applications in economics and across the social sciences. These include case studies, such as text-as-data methods and natural language processing for analyzing policy documents or public sentiment, machine decision making in private and public sectors.

Education

MSc programme in Economics – elective course

MSc in Social Data Science - elective course

 

The PhD Programme in Economics at the Department of Economics:

  • The course is an elective course with research module. In order to register for the research module and to be able to write the research assignment, the PhD students must contact the study administration AND the lecturer.

 

The course is open to:

  • Exchange and Guest students from abroad
  • Credit students from Danish Universities
  • Open University students

 

Full-degree students enrolled at the Faculty of Social Science, UCPH 

  • Master Programme in Social Data Science
  • Master Programmes in Political Science and Social Science
  • Master Programme in Security Risk Management
  • Master Programme in Global Development
Learning outcome

After completing the course, the student is expected to be able to:

 

Knowledge:

  • Understand foundational and advanced machine learning models and their applications to economic and social questions.
  • Reflect on the societal implications of machine decisions and governance challenges across different sectors.
  • Identify and account for fairness aspects in the design and evaluation of prediction algorithms.
  • Analyze the trade-offs and risks of AI implementation, including bias, unintended consequences, and systemic risks.

 

Skills:

  • Apply advanced ML techniques (e.g., causal forests, transformers) to solve real-world problems.
  • Use machine learning models to improve econometric estimation and causal inference.
  • Develop explainable AI workflows for high-stakes environments.
  • Critically evaluate AI-driven policies in terms of societal impact and ethical considerations.

 

Competences:

  • Integrate theoretical and applied knowledge from economics, social sciences, and computer science to address interdisciplinary challenges.
  • Formulate research questions and propose solutions that combine machine learning with insights into societal impacts and governance.
  • Evaluate the design and implementation of machine-driven decisions to ensure fairness, transparency, and ethical integrity.

Lectures and classes.
Direct Instruction, Experiential Learning, Learning through Projects and Inquiry

The course includes readings from recent and foundational research. A tentative list includes (more references will be added on part III):

  • Athey, S., & Imbens, G. (2016). Recursive partitioning for heterogeneous causal effects. Proceedings of the National Academy of Sciences, 113(27), 7353-7360.
  • Athey, S., & Imbens, G. W. (2019). Machine learning methods that economists should know about. Annual Review of Economics, 11.
  • Belloni, A., Chernozhukov, V., & Hansen, C. (2014). High-dimensional methods and inference on structural and treatment effects. Journal of Economic Perspectives, 28(2), 29-50.
  • Belloni, A., Chernozhukov, V., Fernández-Val, I., & Hansen, C. (2017). Program evaluation and causal inference with high-dimensional data. Econometrica, 85(1), 233-298.
  • Chernozhukov, V., Chetverikov, D., Demirer, M., Duflo, E., Hansen, C., Newey, W. and Robins, J., 2018. Double/debiased machine learning for treatment and structural parameters. The Econometrics Journal.
  • Chernozhukov, V., Hansen, C., & Spindler, M. (2015). Post-selection and post-regularization inference in linear models with many controls and instruments. American Economic Review, 105(5), 486-90.
  • Fudenberg, D., Kleinberg, J., Liang, A., & Mullainathan, S. (2022). Measuring the completeness of economic models. Journal of Political Economy, 130(4), pp.956-990.
  • Hastie, T., Tibshirani, R., & Friedman, J. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction. 2nd ed. New York: Springer.
  • Hartford, J., Lewis, G., Leyton-Brown, K., & Taddy, M. (2017, August). Deep IV: A flexible approach for counterfactual prediction. In Proceedings of the 34th International Conference on Machine Learning-Volume 70 (pp. 1414-1423). JMLR.org.
  • Jurafsky, D. and Martin, J.H., 2023. Speech and Language Processing. 3rd ed. 
  • Liang, A., Lu, J., Mu, X., & Okumura, K. (2021). Algorithm design: A fairness-accuracy frontier. arXiv preprint arXiv:2112.09975.
  • Ludwig, J., & Mullainathan, S. (2024). Machine learning as a tool for hypothesis generation. The Quarterly Journal of Economics, 139(2), pp.751-827.
  • Ludwig, J., Mullainathan, S., & Rambachan, A. (2025). Large language models: An applied econometric framework (No. w33344). National Bureau of Economic Research.
  • Mullainathan, S., & Spiess, J. (2017). Machine learning: An applied econometric approach. Journal of Economic Perspectives, 31(2), 87-106.
  • Rambachan, A. (2024). Identifying prediction mistakes in observational data. The Quarterly Journal of Economics, p.qjae013.
  • Wager, S., & Athey, S. (2018). Estimation and inference of heterogeneous treatment effects using random forests. Journal of the American Statistical Association, 113(523), 1228-1242.
  • Wager, S., Hastie, T., & Efron, B. (2014). Confidence intervals for random forests: The jackknife and the infinitesimal jackknife. The Journal of Machine Learning Research, 15(1), 1625-1651.

Programming skills: Familiarity with Python.
Mathematics: Knowledge of how to apply and use linear algebra and basic mathematical analysis for optimization.

Causal Inference and Policy Evaluation: It is recommended that students have followed an introduction to causal inference / applied econometric policy evaluation or similar.

The course is planned to be offered each year jointly between MSc Economics and MSc Social Data Science.

Oral
Collective
Feedback by final exam (In addition to the grade)
Peer feedback (Students give each other feedback)
ECTS
7,5 ECTS
Type of assessment
On-site written exam, 1 hour for on-site under invigilation
Home assignment, Whole semester for home assignment
Type of assessment details
The exam consists of two parts: a home assignment (research proposal, max 3 pages, addressing a societal or economic issue through machine learning) and a 1-hour on-site written exam at the exam house.

Home Assignment:
• Can be written individually or in groups of up to 3 people.
• In group assignments, it must be specified who has written which sections to allow for individual assessment. This must be done using exam numbers, as the exam is anonymous.
• The home assignment is written during the semester and must be uploaded to digital notes before the on-site written exam.

On-Site Written Exam:
• The home assignment must be included in the overall submission along with the on-site exam assignment.
• The on-site written exam allows the use of written aids only.

Weighting and Validity:
• The home assignment and the on-site written exam are weighted 50/50.
• Both parts must be included in the submission for the exam to be valid.
• If a student forgets to upload the home assignment to digital notes or to include it in the exam submission, the exam submission is invalid, and an exam attempt will be used.
Examination prerequisites

There are no exam prerequisites that the student must meet to participate in the exam.

Aid
Written aids allowed

for the 1 hour on-site exam. No internet access.

 

All aids allowed, for the home assignment.

Use of AI tools is permitted. You must explain how you have used the tools. When text is solely or mainly generated by an AI tool, the tool used must be quoted as a source.

Marking scale
7-point grading scale
Censorship form
No external censorship
Exam period

Exam information:

The examination date can be found in the exam schedule   here

The exact time and place will be available in Digital Exam from the middle of the semester. 

More information about examination, rules, aids etc. at  Master (UK) and  Master (DK).

Re-exam

Same as the ordinary.

 

Reexam information:

The reexamination date/period can be found in the reexam schedule  here

More information in Digital Exam in February. 

More information at   Master (UK) and   Master (DK).  

Criteria for exam assessment

Students are assessed on the extent to which they master the learning outcome for the course.

 

To obtain the top grade “12”, the student must with no or only a few minor weaknesses be able to demonstrate an excellent performance displaying a high level of command of all aspects of the relevant material and can make use of the knowledge, skills and competencies listed in the learning outcomes.

 

To obtain the passing grade “02”, the student must in a satisfactory way be able to demonstrate a minimal acceptable level of the knowledge, skills and competencies listed in the learning outcomes.

Single subject courses (day)

  • Category
  • Hours
  • Lectures
  • 28
  • Preparation
  • 135
  • Exercises
  • 28
  • Exam
  • 15
  • English
  • 206

Kursusinformation

Language
English
Course number
AØKK08446U
ECTS
7,5 ECTS
Programme level
Full Degree Master
Ph.D.
Duration

1 semester

Placement
Autumn
Price

Information about admission and tuition fee:  Master and Exchange Programme, credit students and guest students (Open University)

Studyboard
Department of Economics, Study Council
Contracting department
  • Department of Economics
  • Department of Anthropology
  • Department of Political Science
  • Social Data Science
Contracting faculty
  • Faculty of Social Sciences
Course Coordinators
  • Andreas Bjerre-Nielsen   (3-72737f5184807572843f7c863f757c)
  • Jonas Skjold Raaschou-Pedersen   (14-757a796c7e397d6c6c7e6e737a804b7e7a6f6c7e397680396f76)
  • Stephanie Brandl   (16-8081727d756e7b76723b6f7f6e7b71794d807c716e803b78823b7178)
Teacher

See 'Course Coordinators'

Saved on the 06-05-2025
Tilbage

Er du BA- eller KA-studerende?

Er du bachelor- eller kandidat-studerende, så find dette kursus i kursusbasen for studerende:

Kursusinformation for indskrevne studerende

Københavns Universitet
Nørregade 10
1165 København K
Kontakt:
Videreuddannelse og Livslang Læring
videreuddannelse@adm.ku.dk