Privacy and ethics : understanding the convergences and tensions for the responsible development of machine learning

Principal investigator : Sébastien Gambs (UQAM)

Co-investigators :

  • Ulrich Aïvodji (UQAM)
  • Céline Castets-Renard (University of Ottawa)
  • Ignacio Cofone (McGill University)
  • Aude Marie Marcoux (UQAM)
  • Dominic Martin (UQAM)

Summary :

The success of machine learning models is such that they are now ubiquitous in our society. However, their widespread use also raises serious privacy and ethical issues, especially if their predictions are put into action in domains in which they can significantly affect individuals. Consequently, we have witnessed in recent years several initiatives proposing design principles and guidelines for the responsible development of artificial intelligence.

Nonetheless, very few works have explored the tensions but also convergences that can emerge when addressing jointly the privacy and ethical challenges when designing and deploying machine learning models. In addition, a fundamental open question is to investigate whether the achievements of these different objectives is always a positive sum game. 

Thus, to be able to understand how to best address privacy and ethics responsibly when developing machine learning models, we need to first have a clear view on how these concepts interact with each other in a positive as well as negative manner. The objective of the project is precisely to investigate this question by following an interdisciplinary approach at the crossroads of computer science, law and ethics.

Final report :

OPC_final.pdfTélécharger

In addition of the final report and the restitution workshop, the scientific results of the project have also been disseminated through papers (journal, conference or preprint) focusing on specific intersection at the intersection of privacy and ethics (i.e., privacy and fairness, privacy and explainability, fairwashing).

  1. Privacy and fairness :
    Ulrich Aïvodji, François Bidet, Sébastien Gambs, Rosin Claude Ngueveu, Alain Tapp: Local Data Debiasing for Fairness Based on Generative Adversarial Training. Algorithms 14(3): 87 (2021)
  2. Privacy and explainability :
    Ulrich Aïvodji, Alexandre Bolot, Sébastien Gambs: Model extraction from counterfactual explanations. CoRR abs/2009.01884 (2020)
  3. Fairwashing :
    Ulrich Aïvodji, Hiromi Arai, Sébastien Gambs, Satoshi Hara: Characterizing the risk of fairwashing. Accepted for publication at Neurips 2021.

OPC Funded Project

This project received funding support through the Office of the Privacy Commissioner of Canada’s Contributions Program. The opinions expressed in the summary and report(s) are those of the authors and do not necessarily reflect those of the Office of the Privacy Commissioner of Canada. Summaries have been provided by the project authors. Please note that the projects appear in their language of origin.

This content has been updated on 19 November 2021 at 19 h 17 min.