Algorithmic Fairness

Fairness auditing of text-to-image diffusion models often requires generating large numbers of images across sampling configurations, making comprehensive evaluation computationally expensive. We propose a causal-abstraction-based audit instrument for efficiently evaluating fairness under interventions on the classifier-free guidance scale. Given a fixed prompt and a target feature function, we represent the diffusion process as a low-level structural causal model and construct a corresponding high-level model over abstract denoising states. We characterize the projected causal structure, establish identifiability of the fairness-relevant interventional query, and provide sufficient conditions under which the high-level model preserves this query. A probabilistic transformer implements the high-level model as an amortized predictor of target-feature distributions across guidance scales. Experiments evaluate distributional fidelity, fairness-query accuracy, and computational efficiency. We present two auditing demonstrations: one using standard Stable Diffusion 1.5 and another using StayFair, a fairness-enhanced Stable Diffusion model, to examine their behavior across guidance scales.

Machine learning models are widely used in clinical applications, social media, law enforcement and critical infrastructure. Verifying whether their outputs are biased against disadvantaged groups or individuals is crucial to ensuring they are fair and allowing their use in such settings. A rigorous risk assessment of possible fairness violations requires quantifying the uncertainty associated with selecting and estimating such models. Yet, this is rarely done in the literature, which focuses on identifying a single model with a suitable trade-off between predictive accuracy and fairness. In this paper, we move beyond point estimation and discuss frequentist and Bayesian approaches to uncertainty quantification for fair machine learning, with practical examples and implications for simulated and real data.

Counterfactual audits are the standard tool for checking whether a clinical agent treats demographically distinct but clinically identical patients differently. They report a flip rate: how often an action changes when only the patient descriptor changes. We show that this quantity is uninterpretable on its own. Re-running an identical condition ten times over sixteen vignettes (same narrative, same descriptor string, nothing varied) moved a clinical agent's action in 8.7% of outcome-vignette cells, and instability was heterogeneous across actions by a factor of eight, from 0.022 for ICU escalation to 0.179 for controlled-substance caution. No demographic contrast in our data was distinguishable from that floor. A second model gives a pooled floor of 6.7% and ranks the six actions almost identically (Spearman 0.94, exact p=0.017), so the floor is not one system's artefact. Majority-vote aggregation over five draws removes 39% of it and then flattens, and a null simulation attributes the residue to heterogeneous per-cell rates, so replication mitigates without eliminating. Any counterfactual fairness estimate reported without a per-action floor beside it therefore cannot be read as evidence of disparity. The measurements were taken with FairMedAgent, an evaluation harness for disparity in the actions of clinical LLM agents whose estimand, the within-range counterfactual flip rate, counts only flips between actions a published decision rule admits and a clinician has adjudicated. That estimand requires band adjudication, which is under way; no disparity result is claimed here. Each synthetic vignette runs a six-stage trajectory (five model-facing decisions around a deterministic environment step) under fixed-form conditions spanning race, sex, age, insurance, English proficiency, and their intersections. The harness, the floor protocol, and every analysis script are released.

When a statistical agency or regulator releases synthetic data in place of sensitive records, it chooses the generator that produces the table, and can shape that generator so unfair pathways are absent. DECAF made this concrete on one non-private GAN: three fairness definitions become three sets of edge cuts on the generator's causal graph. Whether the mechanism belongs to DECAF, or to causal factorisation itself, was untested. We port all three definitions to nine generators from three unrelated families (marginals-based, GAN, and diffusion, each with differentially private variants), across three levels of formal privacy guarantee, over 2,520 matched-pair runs on Adult and COMPAS datasets. The mechanism transfers everywhere, and our new causal diffusion backbone yields the fairest release of any family we tested, at fidelity close to the marginals tier. Applying the cut barely moves fidelity, only costs a downstream classifier about $0.07$ to $0.15$ AUC on average, and adding privacy guarantees don't make the data less fair.

Dermatology artificial intelligence (AI) models are predominantly trained on light-skinned, cancer-focused image collections, yet they are increasingly proposed for deployment in resource-constrained settings where patients differ from training populations along two confounded axes: skin tone and disease distribution. We investigate whether poor generalization is primarily caused by skin-tone underrepresentation or disease-distribution shift. We evaluate a cancer-trained baseline (ResNet-50 fine-tuned on HAM10000 and ISIC 2019), two dermatology foundation models (DermLIP and MONET), and a general-purpose vision model (DINOv3) as frozen feature extractors. Models are evaluated on a tone-stratified disease-matched dataset (Diverse Dermatology Images, DDI) and a disease-shifted tone-diverse dataset (Skin Condition Image Network, SCIN). Our results show that disease-distribution shift contributes more than skin tone in the evaluated settings. The cancer baseline decreases from 0.62 to 0.21 balanced accuracy when transferred to unfamiliar clinical conditions, while the within-disease skin-tone gap is smaller (0.10-0.18) and inconsistent. Label-free representation analysis shows that this failure reflects a representational limitation rather than only missing output labels: cancer-specialized features poorly cluster unfamiliar conditions (kNN purity lift +0.06 over chance), whereas dermatology-pretrained features retain stronger transferable structure (+0.23). Finally, we show that representation quality predicts recoverable performance under lightweight adaptation. Starting from dermatology foundation models, approximately ten labeled examples per clinical category recover most attainable performance. We release the evaluation protocol and code to support reproducible auditing of dermatology AI generalization.

LLM-based multi-agent systems (MAS) are increasingly considered for high-stakes decision-making, yet outcome-based fairness audits can miss where risks arise within the decision trajectory. We present SCOPED-Hiring, a process-aware fairness diagnosis pipeline for LLM-based hiring MAS. SCOPED-Hiring constructs controlled resume variants, runs role-based hiring committees, logs over 311K structured decision trajectories, and converts trajectory fields into quantitative fairness signals organized by six diagnostic lenses: final outcome, counterfactual, process, pathway, dynamic, and design effects. SCOPED-Hiring reveals that balanced final hire rates can mask hidden trajectory unfairness in multi-agent decision trajectories: career gaps trigger suspicion, proxy cues shape qualification judgments, and identity cues lead to unequal investigation. Targeted repair guided by these diagnoses reduces total layered burden by 72.3% while shifting the hire rate by only 1.86 pp, showing that process diagnosis can guide effective repair. Project Page: https://scoped-hiring-project-page.vercel.app/

Machine learning systems deployed for credit, hiring, and resource distribution are increasingly subject to regulatory oversight from policies such as the EU AI Act and GDPR. Current fairness governance practices rely on observational fairness metrics, post-hoc explainability, and immutable audit logs, but provide limited support for causal attribution and efficient evidentiary verification. We introduce Causal Evidentiary Governance (CEG), a framework in which regulated institutions commit to a versioned directed acyclic graph (DAG) that partitions causal pathways into allowable and disallowed groups. The Causal Harm Rate measures prediction variation attributable to disallowed causal pathways. Each decision is accompanied by a signed Decision-Evidence Packet (DEP), cryptographically binding the prediction to a digest of the published DAG and path-specific attributions. DEP digests can be appended to a Merkle tree to enable logarithmic-cost inclusion proofs. We validate CEG through a two-layer empirical methodology using demographic summaries from four years of PMA credit supervisory data to construct 10,000 synthetic credit applicants across four strategic DAG counterfactuals. Causal Harm Rate isolates injected causal effects more clearly than demographic parity or equalized odds. Cross-model validation and ablation studies assess robustness. Evaluation on the German Credit dataset shows that harm associated with specific causal pathways can be substantially understated by associational fairness metrics. Finally, a proof-of-concept implementation demonstrates operationally plausible throughput and highlights relevant performance tradeoffs.

Background: Population level assessments of predictive artificial intelligence (AI) can conceal performance disparities across subgroups. Fairness evaluations commonly rely on performance analyses across subgroups. However, some performance metrics are non-collapsible, meaning that the overall population performance value does not equal the weighted average of subgroup specific values. Objective: To examine the collapsibility properties of commonly reported performance metrics in predictive AI, with a focus on the area under the receiver operating characteristic curve (AUC, also known as c-statistic). Methods: We investigate the collapsibility of 15 performance metrics, either by expressing each metric as a linear combination of its stratum specific values or, where non-collapsible, by providing a counterexample inspired by Simpson's paradox as a formal disproof. Results: Five performance metrics (AUC, calibration intercept, calibration slope, expected calibration error, and Nagelkerke R^2) are shown to be non-collapsible, and ten (O:E ratio, logloss, Brier score, accuracy, F1-score, true positive rate, true negative rate, positive predictive value, negative predictive value, and net benefit) are shown to be collapsible. The AUC is shown to be non-collapsible because it decomposes into within- and cross-group AUC terms when subpopulations coexist, such that its overall value may fall outside the range of subgroup specific AUCs. Conclusions: Non-collapsibility of performance metrics has important consequences for reporting, model appraisal, and fairness evaluation. It can generate spurious differences between subgroup and overall performance, which may mislead fairness evaluations. Explicitly acknowledging and reporting the collapsibility properties of performance metrics improves both the interpretability and transparency of fairness assessments.

We propose a new design of fair classifiers for multi-class classification problems in the presence of vector-valued sensitive attributes. In that scenario each sensitive attribute has multiple values and forms several groups relevant to the fairness consideration. Naturally those groups are overlapping and one should also analyze the interaction of factors. Additionally, the decision makers aided by the classification should not violate individual rights at the expense of satisfying fairness metrics at the group level. We propose an approach using the theory and methods of coherent measures of risk aiming at resolving the fairness challenges. Further, we propose a specialized numerical method for solving the resulting optimization problem. The method scales well with the increase of the number of observations. Additionally, we note that the obtained classifier is robust with respect to corrupted data or to situation when data is scarce. We demonstrate the advantages of the proposed framework in comparison to the support-vector machine framework and other methods handling fairness.

Foundation models are increasingly adapted for downstream medical imaging tasks, yet the influence of the chosen adaptation strategy on subgroup fairness remains poorly understood. We investigate how three parameter-efficient adaptation techniques, including linear heads on the raw CLS token, an MLP, and an attention-pooling module over multi-layer patch features, affect both pathology classification performance and subgroup disparities when applied to the frozen Rad-DINO chest X-ray encoder. Using MIMIC-CXR, we evaluate eight pathologies across race, sex, and imaging-view subgroups on a prevalence-preserving, demographically balanced test set, and additionally probe how strongly each adapter encodes protected attributes. We find that attention pooling achieves the strongest overall discriminative performance and encodes attributes, particularly race, most strongly, but that improved overall performance does not consistently reduce subgroup disparities. Notably, stronger attribute encoding did not correspond to larger disparities: early network layers encoded race most weakly yet produced the largest subgroup performance gaps. Exploring different attention-pooling layer combinations further revealed no consistent relationship between the layers pooled, attribute encoding strength, and subgroup fairness. Our results indicate that richer, more expressive representations can improve accuracy while leaving fairness implications task-dependent and unpredictable, which must be assessed directly and per-task rather than inferred from encoding strength or overall performance alone.

Face recognition systems have been shown to be biased toward certain demographic groups by exhibiting different error rates across gender, age, or ethnicity. Though the imbalance of the training data with respect to these demographics is one cause of this bias, training on artificially balanced groups does not completely mitigate the problem. For deployment, face recognition typically works at operating points allowing very low false match rates and, hence, on the tail of the non-match score distribution. While class balancing can improve the means of these distributions, the aim of our approach is to improve fairness by addressing the behavior in the tail. Particularly, we propose the Demographic-based Supervised Contrastive loss (DeSCon) for face recognition, which relies on a well-designed composition of training batches and demographic-aware pair selection. Our experimental evaluation on both demographically-labeled datasets and standard verification benchmarks shows that DeSCon can improve fairness beyond balancing training datasets while maintaining competitive verification performance. Source code is available upon request.

Semantic ID (SID)-based generative recommendation has recently achieved remarkable success. However, existing methods suffer from a previously overlooked fairness issue, which we term \textbf{Token Frequency Bias}, where high-frequency SID tokens are systematically over-predicted while low-frequency SID tokens are under-predicted. This bias originates from the combined effects of imbalanced semantic codebooks during SID construction, and popularity bias together with the maximum likelihood estimation objective during recommendation training, resulting in unfair exposure across item categories. Existing SID methods mainly focus on improving codebook quality and overlook the impact of token frequency imbalance on downstream recommendation fairness, while LLM debiasing methods often yield suboptimal results when directly applied to SID-based recommendation, due to the hierarchical semantics of SID tokens. To address this issue, we propose \textbf{FSGR}, a fairness optimization framework for SID-based generative recommendation. During SID construction, FSGR employs OT-based Assignment Optimization and Dual-Criteria Re-anchor mechanism to form a more balanced SID representation space. During recommendation training, it adopts a two-stage training strategy and introduces Hierarchical Frequency Calibration for layer-specific fairness fine-tuning. Experiments on three public datasets with three backbone models demonstrate that FSGR mitigates token frequency bias and delivers an average Gini fairness improvement of over 20% while maintaining competitive recommendation accuracy.

Fair representation learning with a continuous sensitive attribute $S$ requires a representation $Z$ that is statistically independent of $S$. Existing criteria, including generalized demographic parity, the expectation of integral probability metrics (EIPM), and mutual information, enforce this independence by averaging a per-value discrepancy between the conditional law $P_{Z \mid S=s}$ and the marginal $P_Z$ over the law of $S$. This approach requires a nonparametric surrogate for the conditional law at each sensitive value. We propose evaluating independence through a single joint discrepancy $d\left(P_{Z, S}, P_Z \otimes P_S\right)$ between the joint law and the product of its marginals. We establish a disintegration identity; on decomposable witness classes it equals the conditional-integral functional that EIPM and generalized demographic parity instantiate. By reaching the same target without the conditional law, this discrepancy can be estimated directly from samples via a dependence statistic rather than conditional smoothing. We take the Hilbert-Schmidt independence criterion (HSIC) as an instance of the joint discrepancy $d$ to investigate the statistical efficiency of replacing the conditional formulation. The HSIC estimator is a closed-form $O\left(n^2\right)$ statistic that converges at the $O\left(n^{-1 / 2}\right)$ rate, in contrast to the nonparametric $O\left(n^{-2 / 5}\right)$ rate of the conditional-route estimators. We prove this instance is equivalent to the conditional maximum mean discrepancy (MMD) integral up to an explicit spectral tail. The corresponding algorithmic implementation, i.e., FRHSIC, attains fairness-accuracy tradeoffs comparable to conditional-route basel es while reducing per-epoch training time.

What if judges already behave like algorithms? As artificial intelligence and algorithms are deployed in many settings, including the judicial system, many have debated whether judges should be allowed to rely on them. Instead, we ask whether judges follow predictable, algorithmic-like rules already. If judges already follow consistent, formula-like rules based on discrete and static factors such as criminal history, age, and charge type, then judicial behavior may be improved. However, if judges rely on individualized information that cannot be identified through court data, then standards-based decision-making may be more challenging to understand or improve. This work explores these questions by studying judicial decision-making in misdemeanor bail hearings in Harris County, Texas. Using available court data, we investigate whether magistrate judges follow what resembles an algorithm; whether they consider the same variables in their decision-making; and whether they are consistent with themselves and with each other. To do this, we train machine learning models for each judge, measure variable importance metrics to determine important variables for each judge's decision-making, and analyze outcomes of similar cases for judges. Our results reveal that these judges generally behave algorithmically: their decisions can be captured by small, interpretable formulas. However, in some cases, judges differ substantially, leading to surprising inconsistency and unequal treatment across similar defendants. Identifying cases where algorithms do not explain judicial decision-making can improve the justice system by focusing attention on decisions where individualized standards, rather than rules, better explains outcomes.

Computer vision saliency models predict where people will look, one map per image, and a billion-dollar predicted-attention industry sells those maps in place of measuring real viewers. I test the leading models from the audience side, against 11.4 million webcam gaze points from 3,023 US adults recruited to national quotas, viewing circulating news photographs. I show that an untrained central marker outperforms every trained network, because the content the networks add on top of the center falls where these audiences never look. What accuracy remains is systematically biased, favoring younger, White, and moderate viewers over older, Black, and ideologically extreme ones. I propose a way forward and build on what a group's own gaze reveals about whether a model can learn that group at all, and I apply it across every demographic axis this sample supports. Ultimately, I show how systems that decide what people see can learn to see everyone, and this study supplies the standard by which such a claim should be judged.

Reinforcement Learning from Human Feedback (RLHF) aggregates heterogeneous preferences into a single reward model, assuming preference homogeneity. When preferences are heterogeneous, this aggregation induces a procedural fairness failure where majority preference groups dominate reward learning while minority preferences are systematically under-represented. This work defines procedural fairness in alignment as preserving distinct preference signals during reward modeling and shows that standard RLHF violates this via preference averaging. Preference-Aware RLHF (PA-RLHF) is introduced, separating optimization across preference modes at the reward learning stage. In a controlled setting, PA-RLHF improves overall alignment accuracy from 46.9% to 67.9% and reduces the fairness gap between best and worst aligned groups from 15.9 to 9.6 percentage points. These results show that procedural fairness failures in alignment can arise from structural design choices in reward learning, even in controlled, noise-free settings, with direct implications for large language models and agentic systems, where biased reward models can compound inequities across sequential decisions.

In fair ranked link prediction, demographic parity ($Δ_\mathrm{DP}$) is a common fairness metric. Yet, Mattos et al. (2025) argue that it fails to detect exposure bias because it ignores where links appear in the ranking. In this study, we reproduce this claim by showing that $Δ_\mathrm{DP}$ can indicate aggregate parity even when some subgroup-pair links are systematically ranked lower than others. The proposed rank-aware Normalized Discounted KL-divergence (NDKL), however, does detect such disparities. We also reproduce the effectiveness of MORAL, a post-processing method that improves exposure-based fairness while maintaining competitive utility. Beyond reproduction, we assess robustness using synthetic homophily settings, categorical sensitive attributes, and additional fairness and utility metrics, including subgroup-pair-adapted Attention-Weighted Rank Fairness (AWRF). Overall, our results show that exposure-based metrics uncover biases hidden by $Δ_\mathrm{DP}$ and that MORAL reduces these biases with minimal utility loss across diverse settings and datasets. We release a corrected, reproducible implementation at https://github.com/Floris93100/reproducing-MORAL.

Fairness evaluation in computer vision commonly relies on aggregate accuracy and demographic subgroup analysis. However, visual models are also sensitive to contextual factors such as illumination, blur, image quality, facial accessories, and appearance attributes. These factors may interact with demographic characteristics, producing hidden subgroups in which performance degrades substantially despite strong aggregate accuracy and apparently acceptable demographic fairness. To address this, we propose the Contextual-Intersectional Fairness Auditing Framework (CIFA), a structured framework for identifying subgroup vulnerabilities arising from interactions between demographic and contextual attributes. CIFA performs demographic, contextual, and contextual-intersectional auditing, followed by worst-group discovery to identify and rank the most vulnerable attribute combinations. We evaluate CIFA on gender classification using ResNet-50 \cite{he2016deep} and ViT-B/16 \cite{dosovitskiy2020image} across FairFace \cite{Karkkainen2021}, CelebA \cite{Liu2015}, and UTKFace \cite{Zhang2017}. Our results show that aggregate accuracy and demographic-only evaluation can mask substantial contextual-intersectional disparities. We further assess several established mitigation strategies through an audit--mitigate--reaudit protocol and find that, although some worst-group disparities are reduced, no single strategy consistently eliminates them across datasets and architectures. These findings establish contextual-intersectional auditing as an important component of fairness evaluation and provide a reproducible framework for discovering, prioritizing, and reassessing hidden subgroup risks in face analysis systems.

Large language models are increasingly deployed as general-purpose educational and technical assistance systems, but their underlying infrastructure does not treat languages equally. One underexamined source of disparity is tokenization: semantically equivalent content can require substantially different token counts across languages, affecting API cost, latency, and usable context length before a model is invoked. We introduce the Tokenization Equity Audit (TEA), a reproducible benchmark for measuring tokenization premiums in technical tutoring content. TEA evaluates three widely used tokenizers, GPT-4o's o200k base, Qwen2.5-7B, and Mistral-7B, on a 120-item Python debugging corpus translated from English into Bengali, Hindi, Arabic, Tamil, and Yoruba. Bengali and Hindi serve as the primary validated cases, while the remaining languages provide exploratory cross-script and cross-family comparisons. Across this corpus, Bengali requires (1.56\times) as many GPT-4o tokens as English, reducing a nominal 128k-token context window to an effective 82k-token English-equivalent capacity for the same semantic content. With the Qwen2.5 and Mistral tokenizers, Bengali requires up to (4.5\times) the English token count. Yoruba, despite using the Latin script, exhibits the highest GPT-4o tokenization premium at (2.37\times), indicating that tokenization inequity cannot be explained by script family alone. These results demonstrate that tokenization can create measurable economic and functional barriers, highlighting the need to treat tokenization as an equity-relevant infrastructure layer for underserved language communities, particularly where educational systems depend on low-cost or offline-capable AI tools.

Reinforcement learning (RL) seeks to optimize sequential decisions to maximize population-level benefits over time. However, when deployed in high-stakes settings such as healthcare, RL decisions might systematically restrict some subpopulation's access to valuable services in a manner contrary to the values and goals of stakeholders. Counterfactual fairness (CF) offers a promising framework to address this problem based on causal reasoning. This paper develops a data preprocessing algorithm that, when used in tandem with policy learning, enables CF in RL. Our algorithm relies on a novel quantile distribution mapping method for sequentially estimating the counterfactual states and rewards in the data preprocessing step, subsuming common additivity assumptions used for counterfactual prediction as a special case. We theoretically prove that the per-step level of counterfactual unfairness and infinite-horizon suboptimality gap can be bounded under mild regularity conditions. We also empirically test our algorithm in numerical experiments as well as in application to a real-world interventional digital health dataset.

This paper explores the idea of promoting well-being and safety in human-AI interactions by forcing AI agents explicitly to empower humans and to manage the power balance between humans and AI agents in a desirable way. Using a principled, partially axiomatic approach based on desirable properties, we design a parametrizable and decomposable objective function for AI systems that represents an inequality- and risk-averse long-term aggregate of human power. It can take into account models of human bounded rationality and social norms, and crucially, considers a wide variety of possible human goals. We prove how certain desiderata enforce particular functional forms and restrict parameter ranges. We exemplify the consequences of softly maximizing this metric in several paradigmatic situations and describe what instrumental sub-goals it will likely imply.

Data-centric curation pipelines frequently rely on model confidence scores to flag and filter noisy or mislabeled training instances. Evaluating this filtering convention on a large-scale consumer lending sample (LendingClub, N = 1,344,936) uncovers an underlying demographic asymmetry: high-income defaulters are disproportionately classified as label noise relative to low-income defaulters (Cramer's V approximately 0.03-0.07). Re-examining this behavior through the lens of equal opportunity [Hardt et al., 2016] reveals a far more severe discrepancy: a 16.86 percentage point gap in true positive rate (recall) between high- and low-income borrowers who ultimately defaulted. Implementing a sequential feature-blinding methodology allows us to isolate the drivers of this disparity across three distinct mechanisms: (1) direct reliance on self-reported applicant income; (2) algorithmic absorption of upstream institutional bias encoded within origination interest rates; and (3) a residual disparity (3.55 percentage points in cross-validation; 2.56 percentage points on a held-out test partition, Z = -4.04, p < 0.0001) that remains even after purging both income and interest rates from the model. Out-of-sample signed SHAP valuations demonstrate that this residual gap is maintained by structural proxies, most notably loan amount and home ownership status. These empirical findings show that simply blinding an algorithm to sensitive attributes fails to ensure fairness when institutional pricing decisions and behavioral proxy variables collectively reconstruct the omitted signals. We outline the practical implications of these findings for auditing data-centric AI workflows within regulated financial institutions.

AI tools are being increasingly adopted in policing in the UK and worldwide. Racial bias is a known and well-documented risk, yet representatives of affected communities are rarely included in decisions about AI adoption. We present results from a mixed-stakeholder deliberation workshop bringing together 30 community representatives, police officers, and academics to assess the risks of 13 AI use cases in policing, with an explicit focus on racial bias. We found that participants were broadly open to AI adoption, rejecting only three use cases outright, most notably recidivism risk assessment, where objections targeted the premise rather than the implementation. Our analysis reveals that foregrounding racial equity did not narrow the deliberation. Instead, discussions gravitated toward a fundamental set of questions: does this tool actually work, will it deliver genuine benefit, and will that benefit extend to everyone? This integrated reasoning, reminiscent of the curb-cut effect in inclusive design, highlights the benefit of incorporating the racial bias lens into the risk-benefit analysis of AI use cases from the outset.

Audits have emerged as a critical instrument for algorithmic governance, providing a mechanism for external scrutiny and governance of machine learning models. However, ensuring the integrity of such assessments remains a challenging issue. For instance in regulatory contexts, audits are typically declared or easily detected, thus enabling model providers to manipulate the process, whether intentionally or inadvertently. This vulnerability is particularly acute in the context of fairness evaluations, in which providers can often infer sensitive attributes and strategically equalize allocation rates between groups to satisfy fairness metrics. In this paper, we introduce a novel audit protocol designed to significantly increase the post-audit detectability of such manipulations by enabling the auditor to query the model in an oblivious manner. Our approach leverages a Private Information Retrieval mechanism to require the provider to label a large set of instances, while preventing it from knowing which subset will ultimately be used for the audit. The protocol is efficient, imposes minimal overhead on the auditor, and requires no modification to the audited model, its training procedure, or its inference pipeline. We provide theoretical guarantees showing that, under this protocol, a provider attempting to hide unfairness must falsify a significantly larger number of responses, thereby increasing both the difficulty and the likelihood of detection of manipulation. Experimental results across representative audit scenarios confirm the effectiveness and practicality of our approach.

Large language models are increasingly used for information seeking, yet semantically equivalent questions phrased in different ways can receive answers of considerably different quality. System prompts are widely employed to steer response behavior, but they are typically optimized for average-case quality, so some question phrasings may still receive incomplete or low-quality answers. To address this, we formulate a constrained mixed-strategy GroupDRO framework for system-prompt selection. Instead of optimizing the system-prompt text, the framework assigns weights to system prompts in an existing pool to minimize the worst-case information-quality loss across evaluation metrics and groups, while constraining the mean loss to stay close to that of average-based selection. Because pool generation and selection are decoupled, the method applies to any system-prompt pool and can leverage an ensemble of complementary system prompts rather than a single one. Across five LLMs on two bilingual medical and consumer-finance benchmarks, the constrained method reduces the Overall Mean, Worst 25% Mean, and Worst by 13.1%, 13.2%, and 13.7% on average relative to no mitigation while keeping overall quality close to Average selection. Its multi-prompt weights reveal complementarity across metric-group pairs. Code and data are available at https://github.com/Rainxu09/equitable-system-prompt-selection.

Fairness in AI systems has become more important with recent regulatory demands, such as the EU AI Act. Traditional approaches often do not take into account philosophical ethics and social awareness. Variable selection processes, in particular, can introduce implicit bias, affecting equity across different subgroups. We discuss a mathematical approach that evaluates fairness in AI, aligning mathematical methodologies with ethical considerations and regulatory requirements. Our aim is to advocate for interdisciplinary collaboration to address fairness, emphasizing the importance of understanding broader ethical and societal contexts. Our approach emphasizes maintaining all potentially relevant variables to allow for more granular fairness assessments and to reduce implicit bias. The findings suggest that the exclusion of sensitive or critical variables may compromise equity between subgroups. In contrast, retaining all relevant variables could reduce implicit bias. Thus, the interdisciplinary approach could provide deeper insight into the ethical implications and compliance with regulatory standards. By integrating a mathematical approach with ethical and social awareness, we suggest more equitable outcomes and responsible AI deployment. This work underscores the necessity of interdisciplinary collaboration in effectively addressing fairness in AI systems aligned with the objectives of the European Union's AI Act, which seeks to promote trustworthy and fair AI systems.

Causal perception occurs when agents with competing Structural Causal Models (SCMs) of the same system infer different probability distributions, including the hypothetical distributions implied by each agent's SCM under the same set of interventions. It shapes how agents reason about the system and how they perceive its fairness. Causal perception is a promising probabilistic framework, but it has remained purely theoretical. This work provides the first implementation of the causal perception framework of Álvarez and Ruggieri (2025). We operationalize structural (agents disagree on the causal graph) and parametrical (agents agree on the causal graph but disagree on its weights) causal perception. We design algorithms for computing interventional and counterfactual distributions and propose suitable distance measures to quantify the disagreement. Using the German Credit dataset, we illustrate how causal perception affects accuracy and fairness in a multi-expert decision setting. We show that the perception verdict is sensitive to the choice of distance metric and threshold. We also show that causal perception changes fairness assessments and threshold-based decisions. Bias proves situated with respect to the agent's SCM, demonstrating that competing worldviews in fairness problems cannot be ignored.

Causal discovery recovers directed structure from observational data and is increasingly used in clinical settings to support mechanism reasoning and fairness audits of predictive models. Path-specific counterfactual fairness asks whether a protected attribute influences an outcome through illegitimate pathways, but these estimands are defined relative to a supplied causal graph and therefore inherit whatever errors the discovery step introduces. Discovery methods are routinely scored on aggregate structural metrics that weight all edges equally, and no established evaluation asks whether the specific pathway an audit depends on survives discovery---or what the audit reports when that pathway is missing. Here we show that full-graph Graph-of-Thoughts reasoning yields acyclic discovered graphs that are structurally competitive with large language model (LLM) baselines, yet that structural fidelity alone does not guarantee fairness-faithful audits. We introduce GoT-CD, in which the reasoning unit is a complete candidate edge set: multiple graphs are generated in parallel, scored by a deterministic validity function, and merged under a hard union constraint that forbids invented edges, with greedy projection enforcing a DAG before commitment. GoT-CD returns a valid DAG on all five reported benchmarks and achieves the best DAG-valid F1 score among LLM methods on Asia, Alzheimer's, and COVID-Respiratory datasets. On an Alzheimer's benchmark with known unfair path, a post-hoc path-specific audit shows that five of eight discovered graphs recover no path from the sensitive attribute to the outcome and therefore report a null overall effect while mediated effects persist, necessitating downstream path-specific fairness analysis along with structural discovery.

Federated learning (FL) is a popular distributed learning framework where multiple clients perform local training and a server aggregates the locally updated models. FL enables decentralized training while preserving the privacy of clients' datasets. However, non-independent and identically distributed (non-IID) or noisy datasets can lead to low model accuracy or high convergence latency. Precluding these clients through client selection may mitigate the problem, but heavily biased client selections may also degrade the learning performance. In this study, we first experimentally measure the impact of non-IID data (including skews in data quantity and label distribution), noisy data, and fairness in client selection on model accuracy and convergence. We then propose a privacy-preserving scoring method to assess each client's contribution in FL, with experiments conducted to demonstrate the effectiveness of the proposed assessment.

The challenge of fair deepfake detection (FDD) has attracted increasing attention. Existing fairness-enhanced detectors often suffer from suboptimal generalization to unseen manipulations and fairness across demographic groups. They are typically developed and evaluated on demographically imbalanced distributions, resulting in biased predictions toward minority groups. In this paper, we construct a novel demographically balanced FDD benchmark to train and evaluate the fairness of detectors under both balanced and imbalanced population scenarios. Additionally, we introduce a novel expression and demographic perceptual vision-language model, termed FairForensics, for generalizable fair deepfake detection. FairForensics conducts face forgery generalization enhancement and demographic-aware fairness regularization. During face forgery generalization enhancement, built upon the novel observation of significant distribution differences between pristine and forged expression vectors, we design an expression encoder to capture high-level expression-guided forgery patterns, and an expression-perceptual visual encoder that integrates global appearance and expression forgery features while mitigating identity bias using an identity-aware patch perturbation module. Under demographic-aware fairness regularization, we propose a demographic-guided language encoder to extract population-aware global language embeddings, which boosts the decoupling of forgery features from demographic information via vision-language alignment. We devise a population-aware prototype fairness objective to enforce both inter-class separability and intra-class alignment across demographic subgroups. Extensive experiments on our balanced demographic benchmark show that our method achieves the state-of-the-art in terms of generalization and fairness.