KETU 2265 - Responsible Use of Algorithms in the Public Sector: Case Study of the English 2020 Exam Result Algor

Algorithmic decision-making tools are increasingly used in the public sector. At the end of this course, students will be able to understand the implications of and good practices around their design and implementation. The main questions that will be addressed in the course are: when to build algorithms? For whom? How to ensure their efficacy and monitor their actual impact, especially with regards to inequalities? How may civil society push back against them and how to create fruitful dialog with other actors? It will be structured around the case of the algorithm developed by the English government for A-level results during the 2020 covid-19 crisis, and the backlash that ensued. Each session will explore a particular step of the timeline, from March to September 2020. Other cases will be studied to go deeper into one aspect of designing and implementing algorithms.
Soizic PENICAUD
Séminaire
English
Once during the semester, every student will have to: - Prepare and present a group presentation (chosen at the beginning of the semester) - Prepare and submit at least one and up to two individual written assignments (chosen at the beginning of the semester). For every session, all students will have to prepare for the session through various means: academic papers, government reports, newspaper articles, videos, podcasts (depending on the theme of the session). This preparation should take around 2-3 hours, including writing the short paragraph that counts as “class participation”. After each session, students will also have to write a wrap-up paragraph (estimated time: 15 minutes), counting towards “class participation”.
An interest in digital technology for public policy and the will to go in depth into a case are essential. Curiosity about the social and political implications of data and algorithms is an asset. No technical background required. Additional reading on data science will be provided.
Autumn 2021-2022
35% : in-class simulation presentation (1 per student for the course) - group grade. 35% : individual written work (1 per student for the course), with varying formats based on the session just studied: policy memo, audition report, visual mapping, blog post, etc. Students can choose to do 2 throughout the semester and the best grade will be the final grade. 20% : final home assignment. 10% class participation based on completion of 2 shorts reports before and after every class, to be submitted on a public platform: one paragraph before class underlining a point or raising a question about one of the documents provided for the session and one paragraph after class highlighting 3 things the student found interesting and/or learned during class.
The course will be structured around one case study in England. Other cases and areas will also be explored. Each lesson will roughly be structured as: - Group presentations by students followed by class discussion; - Lecture by the teacher giving feedback and summarizing key points; - When relevant, guest interventions; - Collective synthesis and preparation of the next session.
O'Neil (Cathy), Introduction and Chapter 1 (Bomb parts: What is a model?), Weapons of Math Destruction: how big data increases inequality and threatens democracy, New York, Crown, 2016.
Suresh (Harini), Guttag (John), A Framework for Understanding Unintended Consequences of Machine Learning, FAccT 20: Proceedings of the 2020 Conference on Fairness, Accountability and Transparency of Machine Learning, Barcelona, 2020 (link:https://arxiv.
Office for Statistics Regulation, Ensuring statistical models command public confidence: Learning lessons from the approach to developing models for awarding grades in the UK in 2020 - Executive Summary, 2 March 2021 (link: https://osr.statisticsauthorit
Foxglove, Press Release: U-turn on A-level algorithm, 17 August 2020 (link: https://www.foxglove.org.uk/news/press-release-u-turn-on-a-level-algorithm)
10 Downing Street, PM Boris Johnson addresses school children in England (video), 26 August 2020 (link: https://www.youtube.com/watch?v=Q5BHMh7hvDY&t=668s)