This report from the Montreal AI Ethics Institute (MAIEI) covers the most salient progress in research and reporting over the second half of 2021 in the field of AI ethics. Particular emphasis is placed on an "Analysis of the AI Ecosystem", "Privacy", "Bias", "Social Media and Problematic Information", "AI Design and Governance", "Laws and Regulations", "Trends", and other areas covered in the "Outside the Boxes" section. The two AI spotlights feature application pieces on "Constructing and Deconstructing Gender with AI-Generated Art" as well as "Will an Artificial Intellichef be Cooking Your Next Meal at a Michelin Star Restaurant?". Given MAIEI's mission to democratize AI, submissions from external collaborators have featured, such as pieces on the "Challenges of AI Development in Vietnam: Funding, Talent and Ethics" and using "Representation and Imagination for Preventing AI Harms". The report is a comprehensive overview of what the key issues in the field of AI ethics were in 2021, what trends are emergent, what gaps exist, and a peek into what to expect from the field of AI ethics in 2022. It is a resource for researchers and practitioners alike in the field to set their research and development agendas to make contributions to the field of AI ethics.

After challenging the validity of these assumptions in real-world applications, we propose ways to move forward when they are violated. First, we show that group fairness criteria purely based on statistical properties of observed data are fundamentally limited. Revisiting this limitation from a causal viewpoint we develop a more versatile conceptual framework, causal fairness criteria, and first algorithms to achieve them. We also provide tools to analyze how sensitive a believed-to-be causally fair algorithm is to misspecifications of the causal graph. Second, we overcome the assumption that sensitive data is readily available in practice. To this end we devise protocols based on secure multi-party computation to train, validate, and contest fair decision algorithms without requiring users to disclose their sensitive data or decision makers to disclose their models. Finally, we also accommodate the fact that outcome labels are often only observed when a certain decision has been made. We suggest a paradigm shift away from training predictive models towards directly learning decisions to relax the traditional assumption that labels can always be recorded. The main contribution of this thesis is the development of theoretically substantiated and practically feasible methods to move research on fair machine learning closer to real-world applications.

We develop tools for utilizing correspondence experiments to detect illegal discrimination by individual employers. Employers violate US employment law if their propensity to contact applicants depends on protected characteristics such as race or sex. We establish identification of higher moments of the causal effects of protected characteristics on callback rates as a function of the number of fictitious applications sent to each job ad. These moments are used to bound the fraction of jobs that illegally discriminate. Applying our results to three experimental datasets, we find evidence of significant employer heterogeneity in discriminatory behavior, with the standard deviation of gaps in job-specific callback probabilities across protected groups averaging roughly twice the mean gap. In a recent experiment manipulating racially distinctive names, we estimate that at least 85% of jobs that contact both of two white applications and neither of two black applications are engaged in illegal discrimination. To assess the tradeoff between type I and II errors presented by these patterns, we consider the performance of a series of decision rules for investigating suspicious callback behavior under a simple two-type model that rationalizes the experimental data. Though, in our preferred specification, only 17% of employers are estimated to discriminate on the basis of race, we find that an experiment sending 10 applications to each job would enable accurate detection of 7-10% of discriminators while falsely accusing fewer than 0.2% of non-discriminators. A minimax decision rule acknowledging partial identification of the joint distribution of callback rates yields higher error rates but more investigations than our baseline two-type model. Our results suggest illegal labor market discrimination can be reliably monitored with relatively small modifications to existing audit designs.

Algorithmic decision making process now affects many aspects of our lives. Standard tools for machine learning, such as classification and regression, are subject to the bias in data, and thus direct application of such off-the-shelf tools could lead to a specific group being unfairly discriminated. Removing sensitive attributes of data does not solve this problem because a \textit{disparate impact} can arise when non-sensitive attributes and sensitive attributes are correlated. Here, we study a fair machine learning algorithm that avoids such a disparate impact when making a decision. Inspired by the two-stage least squares method that is widely used in the field of economics, we propose a two-stage algorithm that removes bias in the training data. The proposed algorithm is conceptually simple. Unlike most of existing fair algorithms that are designed for classification tasks, the proposed method is able to (i) deal with regression tasks, (ii) combine explanatory attributes to remove reverse discrimination, and (iii) deal with numerical sensitive attributes. The performance and fairness of the proposed algorithm are evaluated in simulations with synthetic and real-world datasets.