Cette page dédiée présente l’un des thème / domaine stratégique actuellement en discussion dans le cadre de notre Programme de financement de recherche stratégique. L’ensemble des thèmes / domaines en discussion est indiqué sur la page du programme. Chaque page dédiée (y compris celle-ci) peut être à un niveau de détail et de maturité variable. Pour participer à la discussion sur ce thème ou en proposer de nouveaux, veuillez utiliser ce formulaire. Si vous souhaitez être tenu.e au courant des développements autour de ce thème / domaine, inscrivez-vous ci-dessous.
Human-Centered Artificial Intelligence (HCAI): From Algorithm Development to Human Adoption of AI
Description et justification du domaine
This theme is a priority for International (e.g., OECD AI Principles), National (e.g. Canada’s Digital Charter: Trust in a digital world), and other governing bodies.
On campus, we have the scientific expertise (from Ivado’s academic members such as Tech3Lab and Mila and from researchers from HEC, Polytechnique and UofM) to tackle this strategic theme.
Ajout 15/07 : Les questions relatives au rôle de l’humain au sein de l’IA sont primordiales pour assurer non seulement son adoption à grande échelle mais aussi son utilisation responsable. L’écosystème québécois de l’IA se démarque d’ailleurs par son souci à cet égard (e.g. Déclaration de Montréal), ce qui contribue à sa réputation enviable.
Contexte
Domaines connexes : AI, Operations Management & Logistics, Human Factors, Information Systems, Marketing, Human Resources, Decision Science. Trustworthy AI (fairness, explainability, robustness, privacy, and safety)
Mots-clés : Explainable AI, Biases in AI, Ethics, User Experience, Algorithms, ML processes, Knowledge Graph, AI Adoption, Éthique, transparence, principe d’autonomie, Automatisation VS augmentation.
Organisations pertinentes : In addition to traditional granting agencies (NSERC, SSHERC), many Tech3Lab’s industrial partners have expressed a continued interest in this research theme (e.g., CAE, BNC, Desjardins, Sobeys, Deloitte, Blue Yonder, Videotron, Dis-Moi, Radio-Canada, CN, LRDG, DBox), OBVIA.
Personnes pertinentes suggérées durant la consultation :
Les noms suivants ont été proposés par la communauté et les personnes mentionnées ci-dessous ont accepté d’afficher publiquement leur nom. Notez cependant que tous les noms des professeur.e.s (qu’ils soient affichés publiquement ou non sur notre site web) seront transmis au comité conseil pour l’étape d’identification et de sélection des thèmes stratégiques. Notez également que les personnes identifiées durant l’étape de consultation n’ont pas la garantie de recevoir une partie du financement. Cette étape sert avant tout à présenter un panorama du domaine, incluant les personnes pertinentes et non à monter des équipes pour les programmes-cadres.
- Pierre-Majorique Léger
- Sylvain Sénécal
- Robert Pellerin
- Constantinos Coursaris
- Ruxandra Luca
- Muhammad Aljukhadar
- Gregory Vial
- Yu Ma
- Joé T. Martineau
- Hyunhwan « Aiden » Lee
- Jean-François Cordeau
- Francesco Ciari
- Marc Fredette
- Jian Tang
- Wietske Van Osch
- Roman Lukyanenko
- Bogdan Negoita
Programmes-cadres potentiels
How can we make AI more human-centered in order to increase its adoption?
The context and challenge :
Human-AI interactions are nowadays common. Humans interact with AI daily as they accomplish many professional and personal tasks. However, human adoption of AI is far from automatic, successful, and satisfactory. Whether it concerns citizens, employees or consumers, issues such as biases, trust, and satisfaction do affect AI adoption by humans in various contexts. For instance, a planning manager may not trust the results of an optimization logistics engine because of its lack of transparency, resulting in over-riding the engine recommandations. Moreover, citizens may distrust facial recognition systems due to biased AI systems which may misidentify certain people.
In the spirit of IVADO’s Apogée research program entitled “Data Serving Canadians”, and building on recent work (e.g. Sun et al, 2020; Bai et al 2020), research is needed to inform and provide guidelines to AI professionals to keep a human focus over the complete life cycle of AI development. We thus proposed a research program oriented toward to following ambitious question :
How can we make AI more human-centered in order to increase its adoption?
The research program:
In order to meet this challenge, a holistic approach to AI is needed. Thus, the proposed multidisciplinary research program will investigate the complete AI development cycle from its inception to its adoption by humans. By investigating how AI is used by humans, and by keeping the human in the loop in the end to end AI development and adoption continuum, it will be possible to inform AI development (e.g., algorithms, ML processes, knowledge graph improvement) with regards to user behaviors (e.g. AI adoption) and user perceptions (e.g., privacy, security, ethics, value), factors likely to influence the trust and continuous use of AI. In order to do so, a wide range of expertise is needed. Thus, researchers from Management science, AI, OM, IT, Marketing are needed to collaborate in a multi-method and iterative methodological approach in order to propose human-centered AI.
Context specific examples:
- AI-based systems that employees actually use and trust in making production/logistics/purchase decisions
- Voice assistant that consumers rely on to make reservations and buy products based on assistant recommendation
- AI systems that are not influenced by irrelevant personal characteristics of the user.
- Effective communication b/w autonomous vehicles and pedestrians
- AI systems that clearly explain to their users why they arrive at the results (XAI)
Examples of publications in the field of HCAI:
Sun, Jiankun and Zhang, Dennis and Hu, Haoyuan and Van Mieghem, Jan Albert, Predicting Human Discretion to Adjust Algorithmic Prescription: A Large-Scale Field Experiment in Warehouse Operations (April 26, 2020). Available at SSRN: https://ssrn.com/abstract=3355114 or http://dx.doi.org/10.2139/ssrn.3355114
Bai, Bing and Dai, Hengchen and Zhang, Dennis and Zhang, Fuqiang and Hu, Haoyuan, The Impacts of Algorithmic Work Assignment on Fairness Perceptions and Productivity: Evidence from Field Experiments (March 8, 2020). Available at SSRN: https://ssrn.com/abstract=3550887 or http://dx.doi.org/10.2139/ssrn.3550887
Documentation complémentaire
(pas de documentation complémentaire pour le moment)
Historique
23 juin 2021 : Première version
3 juillet 2021 : Ajout de noms de personnes pertinentes.
6 juillet 2021 : Ajout de noms de personnes pertinentes.
13 juillet 2021 : Ajout de noms de personnes pertinentes.
15 juillet 2021 : Suppléments d’info (plusieurs sections).
22 juillet 2021 : Ajout de noms de personnes pertinentes.