With the growing importance of AI governance, numerous high - level frameworks and principles have been articulated by policymakers, institutions, and expert communities to guide the development and application of AI . While such frameworks offer valuable normative orientation, they may not fully capture the practical concerns of those who interact with AI systems in organizational and operational contexts. To address this gap, this study adopts a bottom - up approach to explore how governance - relevant themes are expressed in user discourse. Drawing on over 100,000 user reviews of AI products from G2.com, we apply BERTopic to extract latent themes and identify those most semantically related to AI governance. The analysis reveals a diverse set of governance - relevant topics spanning both technical and non - technical domains. These include concerns across organizational processes -- such as planning, coordination, and communication -- as well as stages of the AI value chain, includ ing deployment infrastructure, data handling, and analytics. The findings show considerable overlap with institutional AI governance and ethics frameworks on issues like privacy and transparency, but also surface overlooked areas such as project management, strategy development, and customer interaction. This highlights the need for more empirically grounded, user - centered approaches to AI governance -- approaches that complement normative models by capturing how governance unfolds in applied settings . By foregrounding how governance is enacted in practice, this study contributes to more inclusive and operationally grounded approaches to AI governance and digital policy.

This paper theoretically investigates and experimentally verifies properties of predictive coding network. Although limited to simple representations, the mathematical analysis has impressed the reviewers. The AC thus recommends acceptance of this work.

The argument is over how best to balance the demands of two huge industries: the tech and creative sectors. More specifically, it's about the fairest way to allow AI developers access to creative content in order to make better AI tools - without undermining the livelihoods of the people who make that content in the first place. What's sparked it is the uninspiringly-titled Data (Use and Access) Bill. This proposed legislation was broadly expected to finish its long journey through parliament this week and sail off into the law books. Instead, it is currently stuck in limbo, ping-ponging between the House of Lords and the House of Commons.

Cautious predictions -- where a machine learning model abstains when uncertain -- are crucial for limiting harmful errors in safety-critical applications. In this work, we identify a novel threat: a dishonest institution can exploit these mechanisms to discriminate or unjustly deny services under the guise of uncertainty. We demonstrate the practicality of this threat by introducing an uncertainty-inducing attack called Mirage, which deliberately reduces confidence in targeted input regions, thereby covertly disadvantaging specific individuals. At the same time, Mirage maintains high predictive performance across all data points. To counter this threat, we propose Confidential Guardian, a framework that analyzes calibration metrics on a reference dataset to detect artificially suppressed confidence. Additionally, it employs zero-knowledge proofs of verified inference to ensure that reported confidence scores genuinely originate from the deployed model. This prevents the provider from fabricating arbitrary model confidence values while protecting the model's proprietary details. Our results confirm that Confidential Guardian effectively prevents the misuse of cautious predictions, providing verifiable assurances that abstention reflects genuine model uncertainty rather than malicious intent.

We propose a novel method, termed the M-learner, for estimating heterogeneous indirect and total treatment effects and identifying relevant subgroups within a mediation framework. The procedure comprises four key steps. First, we compute individual-level conditional average indirect/total treatment effect Second, we construct a distance matrix based on pairwise differences. Third, we apply tSNE to project this matrix into a low-dimensional Euclidean space, followed by K-means clustering to identify subgroup structures. Finally, we calibrate and refine the clusters using a threshold-based procedure to determine the optimal configuration. To the best of our knowledge, this is the first approach specifically designed to capture treatment effect heterogeneity in the presence of mediation. Experimental results validate the robustness and effectiveness of the proposed framework. Application to the real-world Jobs II dataset highlights the broad adaptability and potential applicability of our method.Code is available at https: //anonymous.4open.science/r/M-learner-C4BB.

Human-AI conversation frequently relies on quoting earlier text-"check it with the formula I just highlighted"-yet today's large language models (LLMs) lack an explicit mechanism for locating and exploiting such spans. We formalise the challenge as span-conditioned generation, decomposing each turn into the dialogue history, a set of token-offset quotation spans, and an intent utterance. Building on this abstraction, we introduce a quotation-centric data pipeline that automatically synthesises task-specific dialogues, verifies answer correctness through multi-stage consistency checks, and yields both a heterogeneous training corpus and the first benchmark covering five representative scenarios. To meet the benchmark's zero-overhead and parameter-efficiency requirements, we propose QuAda, a lightweight training-based method that attaches two bottleneck projections to every attention head, dynamically amplifying or suppressing attention to quoted spans at inference time while leaving the prompt unchanged and updating < 2.8% of backbone weights. Experiments across models show that QuAda is suitable for all scenarios and generalises to unseen topics, offering an effective, plug-and-play solution for quotation-aware dialogue.

We introduce a method to reduce refusal rates of large language models (LLMs) on sensitive content without modifying model weights or prompts. Motivated by the observation that refusals in certain models were often preceded by the specific token sequence of a token marking the beginning of the chain-of-thought (CoT) block () followed by a double newline token (\n\n), we investigate the impact of two simple formatting adjustments during generation: suppressing \n\n after and suppressing the end-of-sequence token after the end of the CoT block (). Our method requires no datasets, parameter changes, or training, relying solely on modifying token probabilities during generation. In our experiments with official DeepSeek-R1 distillations, these interventions increased the proportion of substantive answers to sensitive prompts without affecting performance on standard benchmarks. Our findings suggest that refusal behaviors can be circumvented by blocking refusal subspaces at specific points in the generation process.

The increasing use of LLMs as substitutes for humans in ``aligning'' LLMs has raised questions about their ability to replicate human judgments and preferences, especially in ambivalent scenarios where humans disagree. This study examines the biases and limitations of LLMs in three roles: answer generator, judge, and debater. These roles loosely correspond to previously described alignment frameworks: preference alignment (judge) and scalable oversight (debater), with the answer generator reflecting the typical setting with user interactions. We develop a ``no-consensus'' benchmark by curating examples that encompass a variety of a priori ambivalent scenarios, each presenting two possible stances. Our results show that while LLMs can provide nuanced assessments when generating open-ended answers, they tend to take a stance on no-consensus topics when employed as judges or debaters. These findings underscore the necessity for more sophisticated methods for aligning LLMs without human oversight, highlighting that LLMs cannot fully capture human disagreement even on topics where humans themselves are divided.

Federated Learning (FL) has emerged as a critical paradigm for enabling privacy-preserving machine learning, particularly in regulated sectors such as finance and healthcare. However, standard FL strategies often encounter significant operational challenges related to fault tolerance, system resilience against concurrent client and server failures, and the provision of robust, verifiable privacy guarantees essential for handling sensitive data. These deficiencies can lead to training disruptions, data loss, compromised model integrity, and non-compliance with data protection regulations (e.g., GDPR, CCPA). This paper introduces Differentially Private Resilient Temporal Federated Learning (DP-RTFL), an advanced FL framework designed to ensure training continuity, precise state recovery, and strong data privacy. DP-RTFL integrates local Differential Privacy (LDP) at the client level with resilient temporal state management and integrity verification mechanisms, such as hash-based commitments (referred to as Zero-Knowledge Integrity Proofs or ZKIPs in this context). The framework is particularly suited for critical applications like credit risk assessment using sensitive financial data, aiming to be operationally robust, auditable, and scalable for enterprise AI deployments. The implementation of the DP-RTFL framework is available as open-source.

Despite their remarkable achievements and widespread adoption, Multimodal Large Language Models (MLLMs) have revealed significant security vulnerabilities, highlighting the urgent need for robust safety evaluation benchmarks. Existing MLLM safety benchmarks, however, fall short in terms of data quality and coverge, and modal risk combinations, resulting in inflated and contradictory evaluation results, which hinders the discovery and governance of security concerns. Besides, we argue that vulnerabilities to harmful queries and oversensitivity to harmless ones should be considered simultaneously in MLLMs safety evaluation, whereas these were previously considered separately. In this paper, to address these shortcomings, we introduce Unified Safety Benchmarks (USB), which is one of the most comprehensive evaluation benchmarks in MLLM safety. Our benchmark features high-quality queries, extensive risk categories, comprehensive modal combinations, and encompasses both vulnerability and oversensitivity evaluations. From the perspective of two key dimensions: risk categories and modality combinations, we demonstrate that the available benchmarks -- even the union of the vast majority of them -- are far from being truly comprehensive. To bridge this gap, we design a sophisticated data synthesis pipeline that generates extensive, high-quality complementary data addressing previously unexplored aspects. By combining open-source datasets with our synthetic data, our benchmark provides 4 distinct modality combinations for each of the 61 risk sub-categories, covering both English and Chinese across both vulnerability and oversensitivity dimensions.