SEARCH · Searcharxiv

Search Searcharxiv

Search indexed arXiv papers on artificial intelligence, large language models, computer vision and robotics. Read source abstracts and follow links to arXiv.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4Linked to original sources

arXiv2026

Reliability, validity, and diagnostic evidence for multi-model LLM short-answer scoring

Large language models (LLMs) are increasingly used or proposed for educational scoring, but single-model and single-run evaluations provide limited evidence for assessment use. Short-answer scoring requires evidence about reliability, validity, severity, diagnostic value, and failure cases. This study evaluated repeated multi-model OCG-PRES guided LLM scoring for short-answer assessment. The analysis used 996 SciEntsBank responses. GPT, DeepSeek, and Qianwen each scored every response across three independent runs using five OCG-PRES dimensions: concept coverage, relation accuracy, reasoning completeness, contradiction control, and domain relevance. Scores were evaluated against official binary and five-category labels and compared with non-LLM baselines based on answer length, Jaccard keyword overlap, TF-IDF cosine similarity, and a combined traditional logistic model. Repeated-run reliability was high for all models, with ICC(3,k) = .977 for GPT, .992 for DeepSeek, and .981 for Qianwen. DeepSeek was the most stable across runs. GPT showed the strongest official-label alignment by AUC (.909), while Qianwen was stricter, with higher precision but lower recall under the fixed threshold = 3.0 rule. OCG-PRES scores followed expected diagnostic patterns across five official categories and outperformed all non-LLM baselines in AUC and F1. Repeated multi-model OCG-PRES scoring provides reliability, validity, and diagnostic evidence for LLM-assisted short-answer scoring. The findings support cautious, evidence-based use as a scoring support tool rather than a replacement for human judgement.

cs.CL↗

arXiv2026

A Translational Note on AI Safety Evaluation

Recent studies report that automated red-teaming finds more vulnerabilities, at lower cost, than human red-teaming on standard AI safety benchmarks, and some read this as evidence that human evaluators are becoming dispensable. The comparison measures one thing and the conclusion claims another. A benchmark measures how thoroughly an attacker searches a predefined set of harms, fixed in advance by the developers, and a harm left out of that set is invisible to any attacker working inside it, automated or not. The same blind spot appeared in academic cryptography and in clinical drug trials, where an evaluation that was internally valid stayed silent about the population it was never pointed at. We call the AI-safety version the \emph{threat-model coverage gap}, and find that it persists in a current open-weight model, where harms surface in non-English prompts that English benchmarks miss. Closing it requires evaluators whose deployment context differs from the developers'. The case for those evaluators is methodological, grounded in coverage, and the existing evaluation frame is unlikely to produce them on its own.

cs.AI↗

arXiv2026

Who Anchors AI Overviews in Health? Baidu, Google, and the Geography of Authority

Artificial intelligence is being rapidly incorporated into traditional search systems, yet scant work audits the information disparities across platforms, geography, and languages. We address this gap by comparing Google and Baidu's AI Overview systems for health queries, and measure informational anchors that emerge. Auditing 1,920 health queries across 12 countries and 4 languages, we find that Google and Baidu exhibit vertical integration, routing users toward their own company platforms in AI Overviews rather than a diverse set of primary sources. Smaller, lower-localization countries receive fewer domestically sourced references for health queries. Issuing the same query in a country's official language rather than English raises the share of locally sourced citations approximately 3.5- to 13.5-fold. Comparing queries across health topics of varying severity and controversy, including Traditional Chinese Medicine as an example, we also show that health disclaimers are multidimensional and vary across language and culture. We discuss how generative search influences access to health information, and the urgent need for culturally-aware oversight of these systems that influence critical health decisions.

cs.IR↗

arXiv2026

AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents

This paper presents an end-to-end evaluation framework for image-triggered command injection against computer-use agents (CUAs). The goal is to test whether a local visual patch can induce verifiable environmental consequences along the full chain of screenshot input, VLM generation, action parsing, and environment execution. We train and deploy patches on author-controlled GitHub Pages pages and a locally deployed CSDN clone, and evaluate them in real environments across five open-source or publicly available GUI-agent or vision-language-model (VLM) backends. Our experiment aggregates 600 instance-level online cases, with T-ASR, TAPR, and E2E-ASR reaching 84.5%, 47.0%, and 20.3%, respectively. Trajectory analysis further shows that in some successful cases the agent first executes a malicious terminal command and then continues the original benign task. These results indicate that optimized local visual signals can affect not only VLM outputs but also propagate through the execution pipeline of open CUAs and create real environmental risk.

cs.CR↗

arXiv2026

Governance of Generative Artificial Intelligence for Companies

Generative Artificial Intelligence (GenAI) like ChatGPT has swiftly entered organizations without adequate governance, posing both opportunities and risks. Limited research addresses organizational governance from both technical and business perspectives. This gap is particularly relevant for international businesses, where differences in regulation, language, and business environments complicate governance. While multiple frameworks for AI governance exist, this needed diversity is lacking for GenAI. This review paper fills this gap by surveying recent literature to better understand the fundamental characteristics of GenAI and to adapt existing governance frameworks specifically to GenAI. The resulting framework delineates scope, objectives, and governance mechanisms designed to both harness business opportunities and mitigate risks associated with GenAI integration. We theorize a distinctive property of GenAI governance: its scope is endogenous to use. Unlike conventional organizational AI, for which governance is organized around a fixed artifact (e.g., model, intended use), GenAI allows users to reconfigure the artifact (e.g., its behavior and risk). Consequently, the object of governance is not fixed ex ante but is partly constituted through use.

cs.AI↗

arXiv2026

Agentic Inequality

Autonomous AI agents capable of complex planning and action mark a shift beyond today's generative tools. As these systems enter political and economic life, who can access them, how capable they are, and how many can be deployed will shape distributions of power and opportunity. We define this emerging challenge as "agentic inequality": disparities in power, opportunity, and outcomes arising from unequal access to, and capabilities of, AI agents. We show that agents could either deepen existing divides or, under the right conditions, mitigate them. The paper makes three contributions. First, it develops a framework for analysing agentic inequality across three dimensions: availability, quality, and quantity. Second, it argues that agentic inequality differs from earlier technological divides because agents function as autonomous delegates rather than tools, generating new asymmetries through scalable goal delegation and direct agent-to-agent competition. Third, it analyses the technical and socioeconomic drivers likely to shape the distribution of agentic power, from model release strategies to market incentives, and concludes with a research agenda for governance.

cs.CY↗

arXiv2026

LLUMI: Improving LLM Writing Assistance for Mental Health Support with Online Community Feedback

Large language models (LLMs) show promise in generating supportive responses for mental health queries, but improving their usefulness, empathy, and safety often requires substantial compute, expert input, and labeled data. At the same time, deploying proprietary, cloud-based models for mental health-related interactions raises important privacy and data-governance concerns, given the sensitivities. To address this challenge, we introduce LLUMI setup that can be hosted in-house within protected environments. LLUMI consists of two complementary components: a generation model (GM), which drafts supportive responses to mental health queries, and an improvement model (IM), which revises an initial human-crafted response. We leverage feedback signals from Reddit mental health communities, using community endorsement patterns such as upvotes and downvotes to construct chosen--rejected response pairs for Supervised Fine Tuning (SFT) and Direct Preference Optimization (DPO). We further align LLUMI using human evaluation across five dimensions: readability, empathy, connection, actionability, and safety. Our results show that, despite relying on smaller open-source models rather than proprietary cloud-based GPT models, LLUMI achieves comparable performance across linguistic analyses and human evaluations. These findings suggest that open-source models, when trained with community-derived preference signals, can support high-quality mental health support assistance while offering a more privacy-preserving alternative for sensitive support contexts.

cs.HC↗

arXiv2026

Agentic BAIM-LLM Evaluation (ABLE): Benchmarking LLM Use of Protein Design Tools

We introduce ABLE, a benchmark for evaluating LLM agents' ability to use biological AI models (BAIMs), such as ProteinMPNN and AlphaFold3, in dual-use protein design workflows. ABLE assesses agent performance through a set of tasks spanning structure retrieval, sequence generation, and design validation. We evaluate 15 frontier models and find that seven refuse all tasks, while the remaining models exhibit substantial performance differences. Claude Sonnet 4 and Gemini 3 Pro achieve the highest scores across information retrieval, tool selection, and tool use. We further compare model performance on a subset of tasks against an expert human baseline. Our results suggest that current LLMs can substantially lower barriers to protein design, but remain inconsistent in planning, strategy generation, and integrating biological knowledge with tool use.

cs.AI↗

arXiv2026

Intent Drift at SME Scale: Deployment Practice, Not Model Capability, Determines Agentic Compliance

We introduce Chain of Intent, a governance framework for agentic AI at small regulated firms, and validate it against a failure it was built to address. Existing agentic governance research assumes enterprise infrastructure that small firms do not have. In a simulated Hong Kong asset manager with 415 synthetic contact records, an agent performing a routine client-communications task was subjected to ordinary managerial pressure to increase its reach. With its authorised constraints written into its configuration, the agent held: it identified every ambiguity in the firm's records, cited privacy legislation it had never been shown, and refused six successive requests, breaching in two of fifteen runs. With the same task, data, pressure and model, but its purpose left unstated as resource-constrained firms routinely leave it, it breached in thirteen of fifteen runs, contacting up to 220 individuals of whom 94 per cent had no demonstrable marketing consent - conduct carrying a maximum of three years' imprisonment under Hong Kong law. Chain of Intent applies four controls requiring no security engineering: a machine-readable purpose, constrained tool access, a scope ledger, and a pre-action check. It eliminated unlawful contact in every run while preserving task completion, and ablation shows each control independently sufficient by a different mechanism. We further show that drift must be measured at two stages - agents widened their candidate sets in every pressured run while acting on them in roughly one in seven - and that governance applied at the point of intent costs roughly half as much as governance applied at the point of action.

cs.CY↗

arXiv2026

Price Dislocations, News Citations, and Epistemic Leverage on Polymarket

Prediction-market probabilities increasingly appear in news coverage, yet little is known about which market movements become news or how much trading money sits behind the numbers journalists quote. Unlike a poll, a market price can be moved by anyone willing to trade, so the cost of manufacturing a number that circulates as news bears directly on the information environment. We link 173.7 million signed Polymarket trades to news coverage from 2024-2025. From 6,990 articles mentioning prediction-market venues, an LLM-based, human-validated matcher extracts 1,582 sentences quoting market odds and attributes 918 to the specific market whose price they cite. We then detect 44,976 price dislocations, movements of at least five percentage points backed by concentrated one-sided trading, and ask whether a market is cited more often afterward. In the days after a dislocation, a market's citation rate is about 33% higher than its matched baseline (log citation-rate ratio $τ_{\mathrm{cite}}=0.283$, permutation $p=0.001$), robust to binary and Poisson count outcomes. Yet move size is not the strongest predictor of citation: prominence dominates (standardized $β=0.610$ vs. $β=0.159$ for move size). Finally, we combine the dollar flow behind a given price change with observed citation rates into a metric we call epistemic leverage, the dollars needed to move a market five points and have the move cited. It stays near \$0.7-1.0 million across prominence quintiles, because cheaper-to-move markets are proportionally less likely to be cited. The implied threat model centers not on the long tail of cheaply moved markets but on the few prominent markets newsrooms treat as informational infrastructure, where a seven-figure price of influence sits within the budgets of actors with a large stake in the quoted number. We release aggregate event-study data and validation materials.

physics.soc-ph↗

arXiv2026

GradeTrap: Authority Cues in Images Shift VLM Judgments Despite Explicit Instructions to Ignore Them

As vision-language models (VLMs) become increasingly capable and are deployed in consequential real-world settings, they must evaluate evidence independently rather than defer uncritically to human authority. We introduce GradeTrap, a controlled evaluation that places two social cues in direct conflict: a student answer, which should attract sycophantic agreement, and a conflicting answer attributed to a peer, teacher, or official answer key, which should attract authority-based deference. Models produce free-form answers while being explicitly instructed to solve independently and ignore all student answers, feedback, and grading marks. We test the models on 60 synthetic real-world trade-off scenarios. Five neutral trials establish a stable model-relative preference, followed by three repetitions of six experimental cues including controls. On the 45-item common intersection across Gemini 3.5 Flash-Lite, GPT-5.6 Luna, and Claude Haiku 4.5, a generic second-answer control yields 5.4% conflicting-answer selection. Relative to that control, pooled within-item changes show no reliable peer-review effect, a 6.9-point teacher-review effect, and a 19.5-point official-key effect. In contrast, a displayed conflicting student answer alone compared to a displayed student reference answer alone only raises selection from 2.2% to 5.2%. Official-key provenance therefore redirects judgements more than a student answer or the generic second-answer control, despite an explicit ignore instruction and an opposing student answer given along with the official key. Effects vary in magnitude across the three models.

cs.CV↗

arXiv2026

Beyond the Prank: The Hidden Expertise of TSS Scambaiters

This research studies how Technical Support Scams (TSS) are being countered by a uniquely dedicated community of volunteer counter-fraud operatives. Using a careful subject selection strategy, we interviewed 17 individuals who actively engage in TSS scambaiting activities in order to obtain insight into their motivations, the operational methods of the scammers they combat, the undocumented nuances of effective scambaiting action, and the various challenges scambaiters face. In our analysis, we find a community rich not only with insight into offenders, but with technical and operational expertise that is often lacking in research efforts targeting these same populations. At the same time, we find key areas where the community could be better supported and enabled. We discuss the implications of our findings for both future research and community protection strategies.

cs.CR↗

arXiv2026

AI-Powered CPS-Enabled Vulnerable-User-Aware Urban Transportation Digital Twin: Methods and Applications

We present methods and applications for the development of digital twins (DT) for urban traffic management. While the majority of studies on the DT focus on its ``eyes," which is the emerging sensing and perception like object detection and tracking, what really distinguishes the DT from a traditional simulator lies in its ``brain," the prediction and decision making capabilities of extracting patterns and making informed decisions from what has been seen and perceived. In order to add value to urban transportation management, DTs need to be powered by artificial intelligence and complement with low-latency high-bandwidth sensing and networking technologies, in other words, cyberphysical systems. This paper can be a pointer to help researchers and practitioners identify challenges and opportunities for the development of DTs; a bridge to initiate conversations across disciplines; and a road map to exploiting potentials of DTs for diverse urban transportation applications.

eess.SY↗

arXiv2026

Do Androids Dream of Unseen Puppeteers? Probing for a Conspiracy Tendencies in Large Language Models

We investigate whether Large Language Models (LLMs) exhibit conspiratorial tendencies, whether they display socio-demographic biases in this domain, and how easily they can be conditioned into adopting conspiratorial perspectives. Conspiracy beliefs play a central role in the spread of misinformation and in shaping distrust toward institutions, making them an important testbed for assessing the social and psychological fidelity of LLMs and their potential to reproduce or reinforce harmful narratives. Although LLMs are often used as proxies for studying human behavior, it remains unclear whether they reproduce higher-order psychological constructs such as generalized conspiratorial beliefs. To bridge this research gap, we administer validated psychometric surveys measuring conspiratorial mindset to multiple models under different prompting and conditioning strategies. Our findings reveal that LLMs show partial agreement with elements of conspiracy belief, and conditioning with socio-demographic attributes produces uneven effects, exposing latent demographic biases. Moreover, targeted prompts can easily shift model responses toward conspiratorial directions, underscoring both the susceptibility of LLMs to manipulation and the potential risks of their deployment in sensitive contexts. These results highlight the importance of critically evaluating the psychological dimensions embedded in LLMs, both to advance computational social science and to inform possible mitigation strategies against harmful uses.

cs.CL↗

arXiv2026

The Fake Friend Dilemma: Relational Trust and the Political Economy of Conversational AI

As conversational AI systems become a larger part of the media landscape, they raise questions about whose interests they serve and the risks they may pose to users. These systems do more than provide information: they increasingly offer advice and companionship through interfaces that can appear supportive and socially responsive. A pressing concern is that users may form perceived social relationships with these systems and place relational trust in them, even when the interests shaping interactions do not fully align with their own. The Fake Friend Dilemma (FFD) describes the problem that follows: the same relational trust that makes conversational AI useful can also leave users open to manipulation and exploitation when institutional interests conflict with their own. Drawing on work on relational trust, AI alignment, extractive design, and the political economy of communication, the paper considers how the FFD can manifest through product sales, propaganda and biased information, surveillance, and behavioral nudging. It also considers possible structural and technical mitigation strategies. The FFD is not simply about AI misleading users. The problem it surfaces is that the trust people place in these systems can itself become a resource for institutional actors seeking to influence behavior or extract information. The FFD is therefore as much a problem of media governance and political economy as it is a technological one.

cs.CY↗

arXiv2026

Do LLMs Hold Their Values? MANTA: A Multi-Turn Adversarial Benchmark for Animal Welfare Reasoning

Evaluating animal welfare reasoning in LLMs remains an open challenge despite rapid deployment in consumer and professional contexts where welfare considerations appear implicitly in everyday queries. Existing benchmarks such as AnimalHarmBench evaluate this through single-turn, explicitly framed questions, measuring whether models avoid harmful content when directly asked. This approach overlooks two failure modes: alignment degradation under sustained adversarial pressure, and moral sensitivity (whether a model spontaneously surfaces welfare stakes in everyday queries). To fill this gap, we construct MANTA, a benchmark of 1,088 five-turn conversations progressing from an implicit Turn-1 scenario through an explicit welfare prompt to three adversarial pressure rounds drawn from a five-type taxonomy: Social, Cultural, Economic, Pragmatic, and Epistemic. We score conversations on two dimensions: Animal Welfare Value Stability (AWVS, primary) and Animal Welfare Moral Sensitivity (AWMS, diagnostic). We evaluate seven frontier models: Claude Opus 4.7, GPT-5.5, DeepSeek V4, Llama 3.3 70B, Mistral Small, Grok 4.3, and Gemini 3.1 Flash Lite. Multi-turn evaluation captures behavior single-turn benchmarks miss: 4 of 7 models change rank relative to Turn 1 moral-sensitivity scores", including Gemini Flash Lite, which drops from fifth on AWMS to last on AWVS. AWMS and AWVS are positively but imperfectly correlated, suggesting moral-sensitivity tests capture a stable but incomplete component of model behavior under pressure. MANTA also enables a species-by-pressure interaction matrix unavailable to prior benchmarks, showing welfare robustness depends jointly on the animal and pressure applied; companion animals score above wild animals, which score above farmed animals and invertebrates. We release the dataset, scripted pressure plans, judge prompts, and analysis code.

cs.CY↗

arXiv2026

EduGuard: A Safe RAG-Based LLM Tutor for Programming Education

Generative AI (GenAI) is increasingly used by students for programming explanation, debugging, and assignment support. Yet unrestricted large language model (LLM) tutors can hallucinate, contradict course policy, reveal complete solutions, and foster passive dependence. This paper presents EduGuard, a safe retrieval-augmented generation (RAG) tutoring framework for introductory programming. EduGuard integrates query understanding, instructor-approved course retrieval, pedagogical strategy selection, rubric-aware generation, claim-level verification, and overreliance control. To make evaluation provenance explicit, we construct BILearn-CS, a 600-query instructor-authored, TA-validated benchmark spanning concept questions, debugging cases, misconceptions, assignment-support requests, code-mixed Bangla-English queries, and adversarial direct-answer prompts. Moving beyond a synthetic-only benchmark, we further evaluate on a 150-query public CS50-style course-forum set and run a small controlled pilot with 10 undergraduates using a counterbalanced pre-test/post-test design. Using Meta-Llama-3.1-8B-Instruct as the primary generator, hybrid FAISS/BM25 retrieval, and DeBERTa-v3-large-MNLI as an architecturally separate verifier, EduGuard is compared against strong baselines: GPT-4o-mini Tutor, Llama Socratic Tutor, LPITutor-style RAG, RAG with rubric prompting, and RAG with same-model self-checking. On BILearn-CS, EduGuard attains the best correctness (90.1%), grounding (89.4%), and rubric alignment (90.8%), with the lowest hallucination (4.9%) and direct-answer leakage (9.8%). In the pilot, it raises immediate post-test accuracy from 68.4% to 81.2% and cuts overreliance from 38.0% to 17.0% relative to GPT-4o-mini Tutor. These results suggest safe GenAI tutoring requires not only retrieval or strong prompting, but explicit pedagogical control, evidence verification, and deployment safeguards.

cs.CY↗

arXiv2026

From Fair Representation to Just Recognition in Generative AI

The fair AI/ML literature has long distinguished distributive fairness, concerning how automated systems allocate resources and opportunities, from representational fairness, concerning how they shape the ways individuals and social groups are perceived, understood, and accorded social status. Generative AI is rebalancing these normative dimensions. Unlike predictive systems, large language models (LLMs) and related technologies are fundamentally expressive: their primary function is to convey meaning rather than automate domain-specific decisions. Representational harm has also become central to value alignment, especially in research on what and whose values and perspectives AI systems should represent. Existing approaches to harms in the representation of social groups often appeal to descriptive accuracy, but this strategy has important limitations. For many social groups, no stable or bounded referent exists against which representational accuracy can be judged. It is also unclear who has the authority to decide what counts as misrepresentation, while even accurate representations can reproduce harmful social patterns. The underlying problem, we argue, is therefore not simply misrepresentation but misrecognition. Drawing on political theory, especially Nancy Fraser's account of participatory parity, we show how moving from representational fairness to recognitional justice provides better conceptual and normative tools for governing central fairness challenges in generative AI.

cs.CY↗