Over the last ten years the literature in computer science and philosophy has formulated different criteria of algorithmic fairness. One of the most discussed, classification parity, requires that the erroneous classifications of a predictive algorithm occur with equal frequency for groups picked out by protected characteristics. Despite its intuitive appeal, classification parity has come under attack. Multiple scenarios can be imagined in which - intuitively - a predictive algorithm does not treat any individual unfairly, and yet classification parity is violated. To make progress, we turn to a related principle, equal protection, originally developed in the context of criminal justice. Key to equal protection is equalizing the risks of erroneous classifications (in a sense to be specified) as opposed to equalizing the rates of erroneous classifications. We show that equal protection avoids many of the counterexamples to classification parity, but also fails to model our moral intuitions in a number of common scenarios, for example, when the predictor is causally downstream relative to the protected characteristic. To address these difficulties, we defend a novel principle, causal equal protection, that models the fair allocation of the risks of erroneous classification through the lenses of causality.

Large Language Models (LLMs) have recently demonstrated remarkable reasoning ability as in Chain-of-thought prompting, but faithful multi-step reasoning remains a challenge. We specifically focus on backward chaining, where the query is recursively decomposed using logical rules until proven. To address the limitations of current backward chaining implementations, we propose SymBa (Symbolic Backward Chaining). In SymBa, the symbolic top-down solver controls the entire proof process and the LLM is called to generate a single reasoning step only when the solver encounters a dead end. By this novel solver-LLM integration, while being able to produce an interpretable, structured proof, SymBa achieves significant improvement in performance, proof faithfulness, and efficiency in diverse multi-step reasoning benchmarks (ProofWriter, Birds-Electricity, GSM8k, CLUTRR-TF, ECtHR Article 6) compared to backward chaining baselines.

Recently, numerous new benchmarks have been established to evaluate the performance of large language models (LLMs) via either computing a holistic score or employing another LLM as a judge. However, these approaches suffer from data leakage due to the open access of the benchmark and inflexible evaluation process. To address this issue, we introduce $\textbf{TreeEval}$, a benchmark-free evaluation method for LLMs that let a high-performance LLM host an irreproducible evaluation session and essentially avoids the data leakage. Moreover, this LLM performs as an examiner to raise up a series of questions under a topic with a tree planing strategy, which considers the current evaluation status to decide the next question generation and ensures the completeness and efficiency of the evaluation process. We evaluate $6$ models of different parameter sizes, including $7$B, $13$B, and $33$B, and ultimately achieved the highest correlation coefficient with AlpacaEval2.0 using only around $45$ questions. We also conduct more analysis to show the robustness and reliability of TreeEval. Our code can be accessed via the provided https://github.com/Ashura5/TreeEval.

In the book "Ethics in Design and Communication: Critical Perspectives," designer and researcher Sarah Edmands Martin wrote that Tinder's design, which presents users with profile cards of potential matches stacked on top of one another, means users "are urged onward" to the next profile "peeking from below the current card, subtly pressuring a user to move on."

The emerging success of large language models (LLMs) heavily relies on collecting abundant training data from external (untrusted) sources. Despite substantial efforts devoted to data cleaning and curation, well-constructed LLMs have been reported to suffer from copyright infringement, data poisoning, and/or privacy violations, which would impede practical deployment of LLMs. In this study, we propose a simple and easily implementable method for purifying LLMs from the negative effects caused by uncurated data, namely, through ensembling LLMs with benign and small language models (SLMs). Aside from theoretical guarantees, we perform comprehensive experiments to empirically confirm the efficacy of ensembling LLMs with SLMs, which can effectively preserve the performance of LLMs while mitigating issues such as copyright infringement, data poisoning, and privacy violations.

AI systems may have transformative and long-term effects on individuals and society. To manage these impacts responsibly and direct the development of AI systems toward optimal public benefit, considerations of AI ethics and governance must be a first priority. In this workbook, we introduce and describe our PBG Framework, a multi-tiered governance model that enables project teams to integrate ethical values and practical principles into their innovation practices and to have clear mechanisms for demonstrating and documenting this.

The sustainability of AI systems depends on the capacity of project teams to proceed with a continuous sensitivity to their potential real-world impacts and transformative effects. Stakeholder Impact Assessments (SIAs) are governance mechanisms that enable this kind of responsiveness. They are tools that create a procedure for, and a means of documenting, the collaborative evaluation and reflective anticipation of the possible harms and benefits of AI innovation projects. SIAs are not one-off governance actions. They require project teams to pay continuous attention to the dynamic and changing character of AI production and use and to the shifting conditions of the real-world environments in which AI technologies are embedded. This workbook is part two of two workbooks on AI Sustainability. It provides a template of the SIA and activities that allow a deeper dive into crucial parts of it. It discusses methods for weighing values and considering trade-offs during the SIA. And, it highlights the need to treat the SIA as an end-to-end process of responsive evaluation and re-assessment.

Sustainable AI projects are continuously responsive to the transformative effects as well as short-, medium-, and long-term impacts on individuals and society that the design, development, and deployment of AI technologies may have. Projects, which centre AI Sustainability, ensure that values-led, collaborative, and anticipatory reflection both guides the assessment of potential social and ethical impacts and steers responsible innovation practices. This workbook is the first part of a pair that provides the concepts and tools needed to put AI Sustainability into practice. It introduces the SUM Values, which help AI project teams to assess the potential societal impacts and ethical permissibility of their projects. It then presents a Stakeholder Engagement Process (SEP), which provides tools to facilitate proportionate engagement of and input from stakeholders with an emphasis on equitable and meaningful participation and positionality awareness.

Reaching consensus on a commonly accepted definition of AI Fairness has long been a central challenge in AI ethics and governance. There is a broad spectrum of views across society on what the concept of fairness means and how it should best be put to practice. In this workbook, we tackle this challenge by exploring how a context-based and society-centred approach to understanding AI Fairness can help project teams better identify, mitigate, and manage the many ways that unfair bias and discrimination can crop up across the AI project workflow. We begin by exploring how, despite the plurality of understandings about the meaning of fairness, priorities of equality and non-discrimination have come to constitute the broadly accepted core of its application as a practical principle. We focus on how these priorities manifest in the form of equal protection from direct and indirect discrimination and from discriminatory harassment. These elements form ethical and legal criteria based upon which instances of unfair bias and discrimination can be identified and mitigated across the AI project workflow. We then take a deeper dive into how the different contexts of the AI project lifecycle give rise to different fairness concerns. This allows us to identify several types of AI Fairness (Data Fairness, Application Fairness, Model Design and Development Fairness, Metric-Based Fairness, System Implementation Fairness, and Ecosystem Fairness) that form the basis of a multi-lens approach to bias identification, mitigation, and management. Building on this, we discuss how to put the principle of AI Fairness into practice across the AI project workflow through Bias Self-Assessment and Bias Risk Management as well as through the documentation of metric-based fairness criteria in a Fairness Position Statement.

Machine learning models are increasingly used in societal applications, yet legal and privacy concerns demand that they very often be kept confidential. Consequently, there is a growing distrust about the fairness properties of these models in the minds of consumers, who are often at the receiving end of model predictions. To this end, we propose FairProof - a system that uses Zero-Knowledge Proofs (a cryptographic primitive) to publicly verify the fairness of a model, while maintaining confidentiality. We also propose a fairness certification algorithm for fully-connected neural networks which is befitting to ZKPs and is used in this system. We implement FairProof in Gnark and demonstrate empirically that our system is practically feasible. Recent usage of ML models in high-stakes societal applications Khandani et al. (2010); Brennan et al. (2009); Datta et al. (2014) has raised serious concerns about their fairness (Angwin et al., 2016; Vigdor, November, 2019; Dastin, October 2018; Wallarchive & Schellmannarchive, June, 2021). As a result, there is growing distrust in the minds of a consumer at the receiving end of ML-based decisions Dwork & Minow (2022). In order to increase consumer trust, there is a need for developing technology that enables public verification of the fairness properties of these models. A major barrier to such verification is that legal and privacy concerns demand that models be kept confidential by organizations. The resulting lack of verifiability can lead to potential misbehavior, such as model swapping, wherein a malicious entity uses different models for different customers leading to unfair behavior. Therefore what is needed is a solution which allows for public verification of the fairness of a model and ensures that the same model is used for every prediction (model uniformity) while maintaining model confidentiality. The canonical approach to evaluating fairness is a statistics-based third-party audit Yadav et al. (2022); Yan & Zhang (2022); Pentyala et al. (2022).