Responsible machine learning
LearningMachine Learning


probabilités et statistiques (niveau M1), bases du Machine Learning

Objectif du cours

This course will examine the mutual relationship between law, computer science and public policy. With the wide spread of artificial intelligence, the link between these domains is becoming more and more ambiguous.

On one hand, the growth in computing capacity combined with the growth in the amount of data collected has revealed the exceptional modeling and data processing capabilities of machine learning. Great advances have been made for a series of specialized tasks such as speech recognition, translation, decision support systems, fault detection, image classification, etc. opening the door to many applications in health, industry, social sciences and many other fields. On the other hand, the automation of certain tasks or decisions raises a number of significant issues, such as the exacerbation of structural inequalities whereby in domains like predictive justice, hiring and finance, algorithms are leading to decisions that benefit mainly those who are already in power and further disadvantage minorities and vulnerable groups. Although automation is often thought of as a “simple” substitution, it reinforces existing biases because it encodes certain norms into complex lines of code that “obscure” the true nature of the problem (Carr, 2017; Dekker, & Woods, 2002, Surveillance Capitalism Zuboff, 2019). In addition, recommendation systems also embody harmful effects such as political destabilization, mental health issues, misinformation and disinformation. The resource intensive nature of the technology has also an effect on the the environment and ecosystems (Strubell et al, 2019). Recent advancements related to generative AI are broadening the gap between the pros and cons of the technology even further, and underscore the need to act towards a common denominator.

Throughout this course, we will debate how balance between the pros and cons of the technology can be achieved. This will be done through unpacking the main pillars of AI, such as fairness, privacy, transparency and accountability, both from a CS perspective and from a legal/ policy perspective. The literature about each one of those components is on the rise both in the CS/ ML domains, and in the social sciences; and even binding laws are starting to enter into force. In each session, we will dive into one of the components, discuss from a theoretical perspective what are the considerations that each domain is adding to the equation, and demonstrate how practically, using real life case studies, we can bridge the gap and merge them into a solution that addresses both. A special focus will be given to the need to address contextual considerations and the implementation of AI in different domains such as healthcare, finance, and criminal justice.

Presentation : here

Organisation des séances

10 séances de 2h
Lieu des cours : faculté de médecine site de Cochin
Numerus clausus : cours limité à 30 places

Mode de validation

  • Critical analysis of reading – 30%: the course is designed around three main blocks, fairness, privacy and transparency. You will be required to choose one of the three topics and within that block to choose one legal and one technical reading and to write a critical analysis about the chosen piece. You can choose whether to write the legal and technical analysis separately or combined, in any case, you should try to link between the two to the extent possible. Your analysis should include a brief summary of the paper, and a critical reflection. If you are writing a separate analysis for the legal and technical papers, each one should be about 800 words, if it is a combined analysis, it should be about 1500 words.
  • 2 technical assignments 30%.
  • Final project, mock trial 40%, in a group, you will be asked to simulate a court trial, you will pick a case study involving automation/ algorithm deployed in a certain domain. The group will be split into 2, where one side will present the plaintiffs (the party bringing forward the lawsuit), and the other side will present the defendant (the party that is being sued). You will be asked to apply the concept that we studied during the semester. The final projects will be presented during a special session that will be held on January 10.
  • Active and meaningful participation would grant you a bonus to the grade
Les intervenants




voir les autres cours du 1er semestre