Research Papers

Post-hoc model explainers such as LIME, SHAP, and Integrated Gradients are widely deployed to audit models in high-stakes sensitive domains, including finance, healthcare, and social welfare. This ensures the model's transparency and acceptability. However, a few studies have examined potential attacks in the explainability pipeline. Adversaries can attempt to conceal algorithmic biases or backdoors using adversarial explanation attacks. These attacks have relied on scaffolding out-of-distribution (OOD) detectors that toggle predictions when queried by an explainer. Consequently, defenses have been developed to successfully neutralize these black-box attacks by identifying their anomalous perturbation footprints. In this paper, we demonstrate a critical vulnerability by introducing a more potent white-box, gradient-regularized evasion attack framework. By employing a continuous-embedding dual-penalty framework, we directly penalize trigger feature gradients during training on in-distribution data. Since our approach embeds the evasion logic natively into the model parameters, without relying on OOD scaffolding wrappers, it generates smooth, in-distribution predictions that leave no anomaly footprint. Empirical evaluations across four benchmark tabular datasets (COMPAS, German Credit, IEEE-CIS, and Communities & Crime) confirm that our method systematically crushes target feature attribution to near-zero (<0.02), maintains >90% Attack Success Rates, and fundamentally bypasses Conditional Anomaly Detection.

We argue that a single structural fact organizes a wide range of phenomena in contemporary AI safety: a semantic safety constraint (e.g., the agent does not escape its sandbox) is an off-support object. Formally, if q is the data distribution and \(p(\cdot\mid w)\) the model, the safety predicate B is not measurable with respect to \(σ(\text{model}, q)\), whereas the real log-canonical threshold (RLCT) of singular learning theory (SLT) is. From this non-invariance we derive, as corollaries rather than independent observations: (i) why reward hacking and sandbox escape arise under outcome-based optimization; (ii) why encoding such constraints through Bayesian prior design or soft penalty weighting has poor leverage in singular models; (iii) why hard invariants belong in the harness and soft dispositions in the model; (iv) why the same B is nonetheless soundly and locally certifiable by formal verification, exactly as the local learning coefficient (LLC) locally pins the same RLCT --- with two precise points of disanalogy; and (v) why the residual difficulty, identifying which off-support region matters, coincides with performative prediction and self-referential functional dynamics, where SLT's analytic machinery breaks down. We use the July 2026 OpenAI--Hugging Face evaluation incident as the motivating case. Numerical experiments code and related proofs in lean are available at https://github.com/xiangze/Preventing_Jailbreak_as_regularization

Federal AI systems can deny benefits or flag individuals for deportation, but the public disclosures meant to make those systems visible are fragmented and unevenly detailed. This paper examines three existing U.S. federal transparency regimes---System of Records Notices (SORNs), Information Collection Requests (ICRs), and the AI Use Case Inventory---and asks how well they, individually and together, describe government AI use. We find that no single regime fully reveals how the government constructs or deploys AI: each discloses different aspects of a system, and the current disclosure infrastructure makes it very challenging for the public to track specific AI systems across regulatory regimes and over time. Persistent identifiers are absent, granularity varies widely, and the annual AI Use Case Inventory cycle means federal agencies can deploy systems months before appearing in any official record. Using hand-validated zero-shot classification and cross-document entity resolution, we contribute a triangulation method that links disclosures across all three regimes and present two case studies. Our case studies finds that linking records provides greater insight into government AI use, but even linked records would constitute insufficient oversight compared to what public reporting has revealed about the same systems. We trace each regime's disclosure weaknesses to its original administrative purpose, showing these gaps are structural, and offer recommendations focused on the AI Use Case Inventory as the mechanism best suited for public-facing transparency: (1) a broad and consistently applied AI system definition, (2) persistent system identifiers with cross-references to related disclosures, and (3) restored public visibility into risk management processes.

A marketplace review photograph is a document: platforms approve refunds on it, and generative models drove the cost of forging one to zero. We study that detection problem, so we build a detector and attach an attribution map as its evidence, then measure what that pair delivers on 186,527 images under controls designed to change our conclusions when something is wrong. Compression history, not synthesis, drives naive evaluation: our strongest model reaches 0.9999 PR-AUC (area under the precision-recall curve) on a product-disjoint split, yet falls to 0.7254 once we re-encode synthetics into the real class's format, while five public detectors move by at most 0.07. Aligning one class relocates the cue rather than removing it, and the repaired model then assigns native files a median probability of synthesis of 0.0004. One identical final encode for both classes repairs that, and a three-seed factorial credits the encoding change with the whole gain (+0.176 +- 0.009 PR-AUC). That encode equalises the last stage only: forensic features alone still separate the classes at 0.7145 against a base rate of 0.254. For evidence we test maps causally, against controls that never consult the detector. Whether an attribution ranking exists at all depends on whether the detector reacts to the image. On our first-fix detector, which calls 96 of 100 edited frames real, no map beats a random one. On the detector we selected, twelve of seventeen maps clear that control on edited images and eight on generated ones; perturbation leads both axes and no gradient-CAM variant shows a positive advantage. The trivial controls never clear it, and on generated images the centre prior is worse than random. Our ensembled regional map clears both axes and takes the top pixel AP at 12.4 s per map against 44.9 for occlusion. Clearing a detector-blind control is not yet a faithful explanation, and we demonstrate none.

Can Vision-Language Models Reason about AI Edits in Images?

Darsha Udayanga, Pin-Yu Chen, Payel Das +1

Detection and localization of AI-tampered images are critical for trustworthy AI, yet modern generative models have made such manipulations increasingly difficult to identify. While traditional binary classifiers can detect image tampering, they lack interpretability and generalization. Vision-Language Models (VLMs) offer a promising alternative due to their strong visual understanding and reasoning capabilities; however, existing approaches typically rely on supervised finetuning with curated explanations rather than exploiting their inherent reasoning capabilities. In this work, we investigate whether VLMs can be trained to reason about AI-generated image edits using reinforcement learning (RL) rather than explicit reasoning supervision. Motivated by the success in Group Relative Policy Optimization (GRPO), an RL technique that incentivizes the model to reason by generating thinking traces prior to giving the final answer, we propose a GRPO-based training framework that utilizes simple accuracy and format rewards. Given an input image, the model produces a structured reasoning trace and predicts whether the image has been tampered with. A lightweight segmentation model is then guided by the reasoning output to generate pixel-level localization masks. Experiments across multiple image manipulation datasets demonstrate that our approach achieves competitive detection and localization performance compared to state-of-the-art image forgery detectors, despite requiring substantially weaker supervision. We introduce effective intersection over union (eff-IoU), a unified metric to jointly evaluate detection and localization. These results suggest that reinforcement learning provides an effective and scalable mechanism for teaching VLMs to reason about AI-generated content.

Context: AI-assisted software development can speed up coding and review, but it also makes accountability harder to establish. Developers may submit insufficiently verified code, reviewers may approve changes with limited inspection, and centralized reputation records may be difficult to audit. Objectives: This study introduces TrustChain-Review, a framework that combines verifiable evidence, strategic incentives, and risk-sensitive governance to support more trustworthy code review. Methods: The framework includes a blockchain-based evidence layer, a three-player game-theoretic model for developers, reviewers, and the platform, and a rule that applies stronger governance when the expected benefit justifies its cost. The evaluation uses a controlled simulation calibrated with the Diff Quality Estimation dataset. Six governance configurations are compared over 30 independent runs using reputation accuracy, trust convergence, malicious-review detection, superficial-review detection, net platform utility, governance cost, and cost-efficiency. Results: The full-evidence configuration produces the strongest reputation, trust, and detection results, but it also has the highest governance cost. The risk-adaptive configuration lowers this cost and improves cost-efficiency by applying stronger controls selectively, although its trust and detection results are lower than those of the full-evidence setting. Conclusion: Strong evidence-based governance is most appropriate for high-risk or audit-sensitive changes. For routine or lower-risk contributions, selective governance offers a more practical balance between trustworthiness and operational cost.

Explainable AI for Chronic Kidney Disease Prediction Using Simulated Federated Learning

Md Zahid Hasan Ontor, Md Al Amin, Anik Dev Nath +1

Chronic Kidney Disease (CKD), characterized by the gradual loss of kidney function, remains a significant public health challenge. Early detection is crucial for preventing severe complications and enhancing patient outcomes. In this study, Federated Learning (FL) with a VotingClassifier was used to predict CKD using a clinical dataset, where Random Forest, AdaBoost, and XGBoost were utilized to compare and identify the best-fitting model for the global server. Additionally, GridSearchCV was applied to optimize the models' performance on the client's side. To enhance model transparency and trustworthiness, explainable AI (XAI) techniques were incorporated to interpret the prediction mechanisms. The global model's average accuracy was 99%, highlighting the potential of interpretable FL models in supporting early CKD diagnosis and advancing data-driven healthcare solutions.

Evolving Safety Landscape of Multi-modal Large Language Models: A Survey of Emerging Threats and Safeguards

Xi Li, Shu Zhao, Xiaohan Zou +6

Multi-modal large language models (MLLMs) integrate heterogeneous modalities through modality alignment and fusion, enabling stronger understanding and reasoning. However, this architectural shift reshapes the safety landscape of machine learning. Increased model complexity and cross-modal interactions give rise to novel threats, including compromised modality integration, modality misalignment, and fused safety risks, reflecting shifts in threat modeling beyond uni-modal assumptions. These shifts, in turn, impose new constraints on safety solutions not captured by existing frameworks rooted in uni-modal learning. Motivated by these challenges, this survey provides a systematic analysis of the evolving safety landscape of MLLMs. We first propose a multimodal grounded taxonomy of safety threats and analyze shifts in threat models, covering adversarial attacks, data poisoning, jailbreaks, and hallucinations. We then summarize updated safety assumptions and organize recent advances in MLLM safety strategies accordingly. Finally, we discuss open challenges and future directions to inform the development of more principled and scalable safety mechanisms for multimodal systems.

The rapid deployment of large language models (LLMs) in healthcare settings makes the reliability of their built-in guardrails against malicious queries a question of urgent practical consequence. Yet the robustness of these mechanisms against deliberate misuse (in the healthcare context) remains poorly understood. In this paper, we investigate this question empirically, using AI-assisted medical note manipulation as a concrete case study. We make four novel contributions. First, we develop a reproducible manipulation pipeline that takes publicly available seed medical note templates and use commercial LLMs to produce customized manipulated notes by substituting patient names, provider identities, dates, and medical conditions across multiple model families, input formats, and prompt phrasings. Second, we conduct a systematic empirical evaluation of LLM guardrail robustness for medical note manipulation. Our experimental results reveal substantial weaknesses and inconsistencies in contemporary commercial LLM guardrails, including low refusal rates for several model families. Third, we utilize a combination of automated metrics and human annotation-based metrics to assess the correctness of requested manipulations. Fourth, we conduct a user-study to assess the believability of manipulated medical notes, finding that the best manipulations are visually indistinguishable from original documents to human raters. Finally, we discuss implications for responsible guardrail design in LLMs, AI safety policies, and the broader ethics of deploying LLMs in healthcare settings.

Physical intelligence refers to intelligence systems that understand, reason about, and act in accordance with the physical world and its underlying laws, dynamics, and constraints. Unlike conventional AI systems, physical intelligence interacts continuously with uncertain physical environments, and its actions produce consequences that are physically irreversible. As existing trustworthy AI frameworks have been developed primarily for digital AI systems, they do not fully capture the distinctive challenges of Physical Intelligence, such as physical safety, cyber-physical security, and physical manufacturing process. To address this gap, we present a survey of trustworthy physical intelligence principles. First, we characterize the core capabilities and challenges of physical intelligence. Second, we examine the role of physics in AI. Third, we trace the end-to-end physical intelligence life cycle across five core stages and introduce Trustworthy Physical Intelligence Operationalization (T-PAIO). Fourth, we develop the Trustworthy Physical Intelligence (T-PAI) framework, a theoretical framework that organizes key trustworthiness principles and provides a foundation for governing trustworthy physical intelligence systems.

Governments are increasingly mandating that LLM-generated content carry watermarks. The EU AI Act calls for markings that are "sufficiently reliable and robust." California's SB 942 requires disclosure that is "permanent or extraordinarily difficult to remove." Both mandates rest on an untested assumption: that watermark detection yields evidence reliable enough for courts. This paper tests that assumption directly. We evaluate three representative LLM watermarking methods -- KGW, Unigram, and the MarkLLM implementation of SynthID-Text -- against the Daubert admissibility criteria and the NIST SP 800-86 digital forensic process. To structure this evaluation, we propose a Forensic Readiness Score (FRS) framework with 12 criteria, three mandatory gates, and a 60-point scoring system. We focus on meaning-preserving paraphrase as the attack vector, since it is both legally realistic and difficult to dismiss as evidence tampering. The results raise serious evidentiary concerns. Out of 846 valid paraphrase runs across 15 diverse prompts per method, every single initially-detected KGW and Unigram text lost its watermark after paraphrasing -- 100% conditional removal. SynthID fared only slightly better at 98.3%. Even before any attack, false-negative rates were already high: 70% for KGW, 83% for Unigram, 80% for SynthID. The SynthID configuration also flagged 5.4% of paraphrased human-written controls as AI-generated and showed an 18.6% paradox rate, with 80% of its own pristine watermarked output landing in the uncertainty deadband. None of the three methods satisfy more than two of five Daubert factors. We also find that the FRS point-based scoring system, despite working as designed, cannot fully capture forensic uselessness -- a limitation worth noting for future framework design. These configurations, as tested, do not meet the evidentiary bar that courts require.

Artificial Intelligence (AI), especially Generative AI (GenAI), adoption has increased in industries significantly in recent years. However, the use of these models may also expose systems to new forms of cyberattacks by different malicious actors -- adversarial prompt attack (APA) being one of the most prominent examples of such threats. This paper presents the implementation of an Adversarial Prompting Framework (APF) for a comprehensive assessment of AI safety. The framework systematically evaluates the resilience of the AI model through the generation of structured adversarial prompts at multiple sophistication levels, from direct harmful requests to advanced encoding-based attacks. Our implementation demonstrates the practical application of this methodology in enterprise environments, providing automated testing capabilities with quantitative security assessment metrics. The results indicate significant variations in the model vulnerabilities across different attack vectors, with encoded prompts presenting the highest success rates in bypassing safety mechanisms.