Research Papers

As large language models (LLMs) are increasingly used to generate text, their misuse has become a growing societal problem, with implications for academic misconduct, model distillation, and cyberattacks. For these reasons, Model Authorship Attribution (MAA), which identifies the source model of a given text, is important for promoting the responsible use of LLMs and improving AI safety. In practice, attribution systems must remain reliable beyond the distributions observed during training. They may encounter texts from unseen domains and writing settings, as well as generations from future versions of model families that were not available when the system was trained. However, existing MAA methods often exhibit substantial performance degradation under such distribution shifts. To address this challenge, we propose Authorship Tracing via Likelihood-based Attribution of Sources (ATLAS), which formulates MAA through source-conditioned text modeling rather than direct source prediction. ATLAS learns how text from each source family is generated and attributes a given text by comparing its likelihood under each learned source condition. We aggregate this likelihood evidence at the token level with clipping so that no single token dominates the decision. We further introduce MAABench, a broad MAA benchmark for evaluating practical OOD settings such as domain and version. Our experiments show that ATLAS consistently outperforms existing baselines across MAABench and OpenTuringBench. In particular, when both the generation domain and the model version change on MAABench, ATLAS achieves 74.27% accuracy, 9.00pp higher than the strongest baseline. Beyond MAA, ATLAS can also help identify distillation source models by measuring which source model a target LLM most closely resembles. Our extended experiments show that ATLAS recovers the known teacher of every distilled model we evaluate.

As artificial intelligence (AI) increasingly supports human decision-making in high-stakes domains, designing effective human-AI collaboration becomes critical. Successful collaboration requires humans to reflect on decision-making processes and adjust their strategies accordingly. Current research typically investigates interventions, such as explainable AI, that target reflection during the decision-making process (reflection-in-action). Less attention has been given to retrospective examination of decisions (reflection-on-action). This work explores reflection-on-action by studying the impact of feedback on reliance behavior and human-AI team performance. Specifically, we investigate feedback interventions targeting overrides of AI advice through a behavioral study (N=244). We find that such feedback influences reflection-on-action, while unidirectional feedback risks reducing subsequent reflection-in-action. Furthermore, we identify opposing shifts in reliance behavior that leave overall human-AI team performance unchanged but show context-dependent effects. Balanced feedback did not show these risks, preserving both concurrent and retrospective reflection, offering insights for designing for appropriate reliance on AI.

This pictorial presents The Body in Algorithmic Colonialism, a screen-based interactive installation that critically examines Emotion-AI through artistic practice. The work uses facial tracking to simulate how commercial Emotion-AI systems translate facial movement into emotional categories and verdicts of compliance or non-compliance. Rather than treating explainability as technical transparency alone, the installation approaches XAI as an embodied encounter with misrecognition: viewers experience how the body is measured, classified, and returned to itself through algorithmic thresholds. Drawing on critical AI studies, affect theory, and practice-as-research, the pictorial positions artistic practice as a method for making affective classification visible and contestable. It documents the conceptual framework, technical process, and critical context of a prototype developed in the first year of a practice-based PhD. No audience responses, interviews, or participant data are analysed in this submission.

Hybrid DeepXAI Fusion Network for Interpretable Big Data Analytics and Secure Cloud Decision-Making Systems

B. Nancharaiah, Ramakrishna Kolikipogu, J. Sulthan Alikhan +3

Big data analytics supports modern intelligent systems, but model complexity and limited transparency can undermine interpretability, trust, and security in decision-making. This paper presents the Hybrid DeepXAI Fusion Network, a deep learning framework that integrates temporal, relational, and contextual features with explainable artificial intelligence (XAI) mechanisms. The architecture combines one-dimensional convolutional neural networks (1D-CNNs) for temporal feature extraction, Transformer encoders for contextual modeling, and graph neural networks (GNNs) for inter-feature dependency analysis. An attention-based fusion mechanism integrates these representations. Class Activation Mapping (CAM) and an explanation loss are incorporated to improve interpretability and explanation consistency. The framework is designed for deployment in a federated cloud architecture, using homomorphic encryption and blockchain-based auditing to support data privacy and accountability. Experimental evaluations across healthcare, finance, and energy datasets report 98.42% accuracy, 97.9% precision, 98.1% recall, and a 98.0% F1-score, outperforming the evaluated CNN-LSTM, Transformer-only, and XGBoost baselines. Under projected gradient descent (PGD) adversarial perturbations, the model reportedly retains accuracy above 91.5%. Scalability experiments report throughput of 80,000 messages per second with latency below 200 milliseconds. These results suggest that the proposed framework has potential for interpretable and secure cloud-based decision-support systems.

Rapid advances in AI capabilities are causing benchmarks to saturate. Once several models approach perfect scores, a benchmark loses its power to differentiate between them, and with it its utility for informing AI safety cases and frontier policies. Yet benchmark items produce far more information than the scores we typically extract from them: a full transcript of how each model solved the task. We investigate whether these transcripts can be used to assess model capabilities even after benchmark scores have saturated. We simulate saturation retrospectively using historical benchmark and model releases, and test whether protocols that access only correctly solved transcripts from saturated benchmarks can predict capabilities measured on held-out, unsaturated items. We evaluate representative contemporary approaches for extracting information from model transcripts: pairwise judging, single-transcript absolute scoring, and predictions from deterministic transcript features or embeddings. We find that saturated transcripts retain capability information beyond benchmark accuracy. Amongst our evaluated protocols, pairwise judging performs best, reducing prediction error by 5.3 times compared to a baseline using only saturated benchmark scores. This capability-relevant information is distributed across transcript sections, but concentrated in specific channels of information such as reasoning and tool-calls. However, the predictive power of protocols does not always generalise across distribution shifts -- for example by underestimating the capabilities of the first reasoning models introduced in early 2025. Saturated benchmarks thus retain capability-predicting signal in their task transcripts, but extending the lifetime of existing benchmarks requires methods designed to extract these signals robustly.

Most debate about releasing capable AI systems treats the choice as binary: keep the model behind an API, or publish the weights. That framing runs together two separate decisions. One is how capable a system may become. The other is who may obtain a given capability, and on what terms. This paper is about the second decision. We propose a Multi-Tiered AI Accessibility Model (MAIAM) that sorts access by purpose, safety, and capability instead of by release artifact. Tier 1, AI for Learning, is broadly accessible and deliberately capability-bounded. Tier 2, AI for Exploration, serves research and development under registration, identity, and usage controls that tighten as capability rises. Tier 3, Full-Scale Frontier AI, is available only through structured, recorded, and accountable access. The principle behind the tiers is that accessibility should be proportional to capability, safety, and risk. We show why publishing frontier-class weights forfeits every safety lever a developer has after release, and, using evidence that safety fine-tuning can be stripped from open weights for a few hundred dollars, why an evaluation of the base model says little about its derivatives. We give criteria for assigning a system to a tier along capability, autonomy, domain-consequence, and usage/safety axes, and combine MAIAM with the SAFE AI Operating Point into a two-part framework: a Capability Boundary that fixes the safety limits not to be crossed, and an Accessibility Boundary that fixes who may obtain a capability below that limit. Capability limits alone do not make AI safe; a tiered access regime administered by regulating bodies is also needed.

Federated fine-tuning of large language models (LLMs) allows multiple clients to collaboratively adapt models without sharing raw data, offering strong privacy benefits. However, this setting is highly vulnerable to backdoor attacks, in which malicious clients manipulate their updates to induce targeted behaviors while maintaining normal performance on benign inputs. In this work, we make a key observation specific to federated LLM fine-tuning: malicious clients tend to exhibit strong consistency in their model updates, whereas benign clients produce more diverse and dispersed updates. This structural difference is not well captured by existing defense methods, which typically rely on direct anomaly detection in high-dimensional parameter space and therefore struggle in the presence of entangled LLM updates. Motivated by this observation, we propose FedGraph, a graph-based aggregation framework that explicitly models relationships among client updates. By representing clients as nodes in a dynamically constructed graph and extracting structural features that characterize their relative positions and interactions, FedGraph derives compact client representations that effectively separate malicious and benign behaviors. These representations are then used to identify and suppress malicious contributions during aggregation. Extensive experiments demonstrate that FedGraph significantly improves backdoor defense. It consistently reduces attack success rates to below 10% while achieving high detection accuracy (95.5% on average) and a low false positive rate (2.33%), substantially outperforming existing federated learning defenses.

Nurse scheduling is a complex optimization problem in which nurses are assigned to shifts to meet staffing demands while complying with working-time limits, qualification requirements, and other operational and regulatory constraints. Artificial intelligence is increasingly being used to support this process by forecasting future staffing needs and providing advice on how to improve shift scheduling. Effective human-AI collaboration in nurse scheduling requires users to understand, interpret, and appropriately act on AI forecasts and advice. However, it remains insufficiently understood how healthcare professionals use transparency mechanisms like uncertainty communication and AI advice explanations when making staffing decisions. Applying a human-centered design approach, we iteratively developed a prototype decision support system for nurse scheduling and evaluated it through a mixed-methods user study with expert users. Using thematic analysis, questionnaire data, and eye tracking, we examine how healthcare professionals interact with the system and how uncertainty communication and explanations of AI advice influence understanding, decision confidence, trust, and usability. We found that the uncertainty communicated in AI forecasts was frequently misinterpreted and reduced decision confidence. Explanations were generally perceived positively, but did not significantly affect decision confidence or advice acceptance. We conclude with eight design implications derived from our findings, providing insights for the design of more transparent and effective AI-supported decision support tools in healthcare.

XAI-Guided Fourier Domain Generalisation for Pure-Simulation Sim-to-Real Deep Reinforcement Learning

Wei Yao, Chuyao Wang, Ziwei Wang +1

Deep reinforcement learning has shown strong potential for vision-based autonomous driving, but policies trained in simulation often suffer severe performance degradation when deployed in the real world. Existing sim-to-real methods commonly address this gap through image-transfer-based domain adaptation or policy-agnostic domain randomisation. However, the former requires laborious acquisition of target-domain data and is often limited to the deployment domain, whereas the latter introduces unguided perturbations that may not reflect the underlying sim-to-real discrepancy, leading to suboptimal transfer performance. To overcome these limitations, we propose XFD, an explainable artificial intelligence (XAI)-guided Fourier domain generalisation framework for pure-simulation sim-to-real reinforcement learning. The proposed framework extracts explainability attribution maps from simulated observations, and converts these attributions into policy-conditioned Fourier-domain interventions. By perturbing appearance-sensitive spectral components while preserving task-defining spatial structures, XFD constructs robustness-enhancing visual domains entirely within simulation. Subsequently, the proposed robust policy optimisation guides the policy towards attending to task-critical features while suppressing sensitivity to irrelevant visual information. Extensive experiments in simulation and zero-shot real-world deployment demonstrate that XFD consistently improves traveled distance and reward over standard policy and randomly domain-randomised baseline, especially on unseen maps with cluttered backgrounds. Attribution-based analysis further proves that the proposed policy preserves more consistent task-relevant visual evidence across simulation and real-world observations.

Existing human-centred XAI research predominantly evaluates explanations after they have been generated. However, this fails to provide actionable insights into improving accessibility. This study incorporates human-based feature understandability into feature selection to improve the generation of human-centred explanations. Natural language explanations for machine learning models trained on tabular data often cite influential features in the input data to explain a given decision. As a consequence, users' ability to understand the produced explanation hinges on their understanding of the cited features. To quantify this natural-language feature understandability, Rossberg et al. proposed the Feature Understandability Scale, which assigns numerical `understandability scores' to features based on user feedback. To evaluate the degree to which these scores can improve machine learning explanation quality, this study collects understandability scores across two datasets, proposes an understandability-aware feature selection algorithm to select features that are both understandable and informative, and trains a random forest based on these feature subsets. The SHAP-based explanations generated by this model are evaluated using user feedback, and changes in explanation quality are assessed. The individual evaluation of the explanations showed that while the understandability-based explanations scored higher on average, the difference was not statistically significant. In contrast, the comparative evaluation found that participants preferred these explanations significantly more frequently than expected at the chance level. It is concluded that the integration of understandability can yield more understandable explanations at minimal cost to classification performance, although changes to quality require elucidation in larger cohorts.

On many estimates, artificial general intelligence (AGI) is likely to emerge within the next few decades or even sooner. Once it does, it may quickly surpass human intelligence, possibly by a huge margin. Some believe that the country that develops AGI first could attain global dominance. This prospect risks setting off intense competition between the United States and China. Indeed, a narrative has already developed that the superpowers are locked in an AI race, with the Trump administration proclaiming its determination to 'dismantle unnecessary regulatory barriers' to winning it. We argue that this approach presents three profound dangers. The first two are familiar from past arms races: being overtaken by a rival or triggering a war. The third is new: losing control of human beings' future to agents far more intelligent than we are, which could endanger our very survival. Striving to get to AGI first risks a race to the bottom on safety practices and regulation concerning them. China does not yet appear bent on racing to superintelligence. The United States should prioritise both unilateral and co-operative efforts to improve AI safety and avoid provoking a race that all humanity could lose.

Social media platforms generate vast volumes of user-generated content that often obscure meaningful discourse, a phenomenon conceptualized as social noise. While existing misinformation detection methods primarily focus on factual verification, they largely overlook the role of social noise in amplifying misinformation. This study assesses the performance of six text-based models and a graph neural network (GNN) in mitigating the effects of misinformation using predefined social noise constructs. The dataset consists of 500,000 tweets collected from the X platform using the Ukraine hashtag during the Russia-Ukraine conflict from March 2024 to March 2025. Advanced machine learning models achieve strong performance, with accuracies ranging from 79.94% to 97.17%. RoBERTa demonstrates the highest performance, with an accuracy of 97.17%, Cohen's kappa of 0.9660, MCC of 0.9660, and AUC of 0.9977. The GNN achieves an accuracy of 64.52% and an AUC of 0.899, demonstrating performance above random-chance classification by exploiting the relational structure between tweets. All text-based models achieve kappa values greater than 0.75, indicating substantial overall agreement. Explainable AI analysis using LIME reveals that 89.6% of predictive weight is attributed to contextual features rather than isolated keywords. The findings highlight that integrating social noise into AI pipelines enhances misinformation detection and supports more effective mitigation strategies in large-scale social media environments.