News & Research
The latest AI research and news with real-world stakes — each item sourced, dated, summarized in plain English, and tagged by impact area. Every item is checked against its source before it appears.
Research
Strategic Adoption of Artificial Intelligence for Cybersecurity in Small and Medium-Sized Enterprises
Peter Anthony Ene, PhD Nsikak Stephen Edet
Zenodo (CERN European Organization for Nuclear Research) · 2026-09-10
This study surveyed 240 small and medium-sized enterprises (SMEs) to examine what drives or hinders their adoption of AI-based cybersecurity tools, using the Technology-Organisation-Environment (TOE) framework. Findings show that technological compatibility, management support, and external regulatory pressure significantly influence adoption, while cost constraints and skills shortages are the most persistent barriers. The authors conclude that AI adoption for cybersecurity in SMEs is a strategic organizational decision—not just a technical one—requiring managerial commitment and policy support. Recommendations are directed at SME operators, policymakers, and institutions building digital and cybersecurity literacy.
- Enterprise
- AI policy
Research
Strategic Adoption of Artificial Intelligence for Cybersecurity in Small and Medium-Sized Enterprises
Peter Anthony Ene, PhD Nsikak Stephen Edet
Zenodo (CERN European Organization for Nuclear Research) · 2026-09-10
This study surveyed 240 small and medium-sized enterprises (SMEs) to examine what drives or hinders their strategic adoption of AI-powered cybersecurity tools, using the Technology-Organisation-Environment (TOE) framework. Chi-square analyses found that technological compatibility, management support, and external regulatory pressure significantly influence adoption, while cost constraints and skills shortages are the most persistent barriers. The paper concludes that AI adoption for SME cybersecurity is fundamentally a strategic organisational decision, not just a technical one, and calls for policy support and digital literacy initiatives to help resource-constrained firms improve cyber-resilience.
- Enterprise
- AI policy
Research
AspisAI: A Canonical, Machine-Interpretable Governance Framework for Automated Multi-Standard Compliance Monitoring
Tsafac Nkombong Regine Cyrille, Hasan Dag, Reiner Creutzburg et al.
arXiv · 2026-09-09
AspisAI is a machine-interpretable governance framework that translates requirements from multiple cybersecurity and privacy standards—including ISO/IEC 27001, NIST CSF 2.0, Cyber Essentials, and GDPR—into a unified, condition-based compliance model. Within a bounded scope of 26 representative requirements, the framework achieved 88.5% mapping coverage, full traceability, and correct detection of all introduced compliance gaps in a controlled simulation. Cross-standard mappings were validated against NIST's published informative references, yielding 57% exact agreement, and the framework was also applied to real third-party evidence from the OpenSSF Scorecard to surface genuine governance gaps in a live open-source project. The work demonstrates that automated, auditable multi-standard compliance monitoring is achievable, reducing reliance on costly manual spreadsheet-based tracking and periodic audits.
- Enterprise
- Quality assurance
- Certifications
- AI policy
Research
Following the Preference, Missing the Optimum: Compliance Without Optimization in AI Housing Recommendation
Hsuan Lo
arXiv · 2026-09-09
This paper audits large language model (LLM) housing recommendation systems against a verified ground truth of 120 real New York City listings per scenario, finding that while models comply with user-stated preferences at near-perfect rates (only 1.8% violation), 39% of recommendations are strictly dominated—meaning a cheaper, faster-commute, equally large listing existed in the same pool. The dominated recommendations were a median $900/month more expensive and 3.5 minutes farther in commute time than available superior alternatives, a pattern that replicated across both OpenAI and Anthropic models. The study identifies this failure mode as 'compliance without optimization,' where models honor stated preferences without actually finding optimal matches, and proposes dominance-rate instrumentation as a practical diagnostic for deployment. These findings matter for consumer protection and policy in high-stakes domains like housing, where AI acting as a first point of contact can systematically steer users away from their best available options.
- AI policy
- Quality assurance
Research
No-Box Vulnerability Analysis: Description-only Detection of Indirect Prompt Injection Vulnerabilities in MCP Servers
Zehua Zhang, Jie Hu, Pratham Hegde et al.
arXiv · 2026-09-09
This paper introduces 'no-box vulnerability analysis,' a new paradigm for detecting security flaws in AI-integrated systems using only publicly available functionality metadata—no system access or runtime interaction required. The authors implement MCPSEC, a prototype that audits Model Context Protocol (MCP) servers for indirect prompt injection vulnerabilities using only tool metadata exposed at server registration. Evaluated on 20 widely deployed MCP servers with 177 tools, MCPSEC achieved 98.9% recall in identifying the 94 out of 95 human-verified vulnerable tools, outperforming an LLM baseline that achieved 84.2% recall. This work matters because it enables third-party security analysts to audit closed-source, remotely hosted, or commercially gated AI tool ecosystems without needing privileged access, lowering barriers to proactive AI security auditing.
- Quality assurance
- AI policy
Research
Beyond Static Guarantees: Measuring the Static-Pass Dynamic-Fail Gap in Security-Sensitive and LLM-Generated Python Code
Jessica Pourleyli, Maitreyee Das Urmi, Glaucia Melo
arXiv (Cornell University) · 2026-09-09
This paper identifies and quantifies the 'Static-Pass Dynamic-Fail' (SPDF) phenomenon, where code that passes static security analysis tools (Bandit and Semgrep) still contains exploitable vulnerabilities at runtime. Using a three-stage agentic pipeline combining static scanning, LLM-based CWE reasoning, and autonomous exploit verification in Docker containers, the authors evaluated 1,355 Python samples from three security-focused datasets. Of the 654 samples that cleared static analysis, roughly 1 in 7 (14.53%) were found to have runtime-confirmed or partially confirmed exploitability, with certain vulnerability classes like CWE-338 and CWE-916 missed entirely by both static tools. The findings demonstrate that static analysis and runtime security are distinct, hierarchical assurance layers — a distinction with direct implications for how AI-generated and security-sensitive code should be evaluated.
- Quality assurance
- Enterprise
Research
Governing AI Research Through Peer Review: A Mixed-Methods Study of the Longitudinal Effects of Ethics Flags Across Resubmissions
Kento Nishi, Alec Laprevotte, Isaiah Bullock et al.
arXiv · 2026-09-09
This mixed-methods study examines whether ethics flags enforced by selective AI conferences (specifically ICLR) actually redirect research toward safer practices. Tracking 446 rejected or withdrawn submissions with ethics flags into later public resubmissions, the authors find that in 83% of cases authors leave flagged concerns unaddressed or revise only the paper's framing without changing the underlying methods or procedures. Qualitative interviews reveal that authors treat peer review as an 'editorial process' shaping presentation rather than a mechanism for altering research direction, often conceding concerns during rebuttal only to drop those concessions after rejection. The authors recommend policy changes—particularly disclosure of prior ethics flags upon resubmission—so that accountability carries over across review cycles.
- AI policy
- Quality assurance
Research
Towards a Deterministic Math Solver for Clinical Language Models
Felipe Ocampo Osorio, Sebastián Andrés Cajas Ordoñez, Maximin Lange et al.
arXiv · 2026-09-09
This paper tests whether large language models used in clinical settings can avoid arithmetic errors by having the model write case-specific Python code that a deterministic local executor runs, rather than performing calculations directly. Evaluated on MedCalc-Bench Verified (1,100 cases, 55 calculators) using two open-weight models (Qwen2.5-7B and Qwen2.5-32B-AWQ), the approach shows meaningful accuracy gains only for the larger 32B model (+7.05 percentage points, confidence interval clear of zero), while gains at 7B are not statistically reliable. The authors also audited benchmark formulas against current clinical guidelines and flagged 16 of 55 calculators with version, usage, or coefficient concerns, underscoring that a deterministic executor cannot substitute for verified formulas or reliable variable extraction.
- Quality assurance
- Certifications
Research
IBIB: A Protocol for Measuring Enterprise AI Systems by Serving Route, Not Model Identifier
Blake Stenstrom, Charangan Vasantharajan, Brian Sathianathan
arXiv (Cornell University) · 2026-09-09
This paper identifies a systematic measurement error in enterprise AI benchmarking: all 18 audited benchmarks score models by their advertised identifier rather than by the actual serving route (which jointly depends on weights, precision, output contract, and harness). The authors propose IB2, a three-part evaluation protocol that binds capability assessment to the specific route being served, includes reliability failures in scores, and uses score-blind adjudication. Testing across eleven systems reveals that capability availability, benchmark discrimination, and scoring conclusions all shift meaningfully depending on serving-arm configuration—for example, one revision's score moved from 77.38 to 82.54 depending on serving arm, and excluding failed responses from denominators changed point orderings entirely. The work argues that enterprises need route-level measurement protocols to accurately assess AI system performance in deployment.
- Enterprise
- Quality assurance
Research
Emergency Department Revisit Quality Review Screening: Exploring Human Decision-Making and Artificial Intelligence Support
Jonathan A. Handler, Marlene I. Robles-Granda, Jacob E. Mefford et al.
arXiv · 2026-09-09
This exploratory retrospective study examined whether AI tools could support quality-assurance screening of emergency department return visits occurring within 1–14 days, a broader window than the 48–72 hours typically reviewed. The study found that GPT-4 alone performed poorly, flagging 94% of diagnosis pairs as warranting follow-up compared to far lower rates among clinician raters, while a knowledge-graph-augmented algorithm (KGA) achieved 83–100% positive predictive value for identifying pairs that at least one clinician deemed concerning. Among human raters, the medical gravity of the revisit was the factor most consistently associated with a case warranting further review. The authors conclude that the KGA approach may allow health systems to expand the scope of revisit screening without substantially increasing reviewer workload, though further validation is needed.
- Quality assurance
Research
Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization
Ayan Majumdar, Shounak Paul, Pushpdeep Singh et al.
arXiv · 2026-09-09
This paper evaluates whether Vision-Language Models (VLMs) can reliably moderate online content by comparing two guidance paradigms: instruction-driven (reasoning from policy rules) and example-driven (generalizing from precedents). The researchers introduce ModerationBench, a new benchmark of 4,000 manually annotated posts from the Bluesky platform, and find that foundation models substantially outperform Bluesky's existing moderation system, nearly tripling its F1 score (0.60 vs. 0.22) on random posts. Both paradigms achieve comparable peak effectiveness, suggesting a path toward scalable and adaptable content policy operationalization. These findings matter for platform policy and quality assurance, as they demonstrate that AI foundation models can address the challenge of consistently applying complex moderation policies at scale.
- AI policy
- Quality assurance
Research
Cyber-Financial Contagion: Modeling the Propagation of an AI Vendor Compromise Through the Banking System
Alex Leytes
arXiv · 2026-09-09
This paper examines how a security compromise at a single AI vendor—such as those providing fraud screening, credit decisioning, or anti-money-laundering tools to banks—can cascade through the broader banking system and trigger losses resembling a classical financial crisis. The authors build a four-layer network model (CFC-Prop) coupling AI vendors, financial institutions, interbank exposures, and customer accounts, and demonstrate on a synthetic dataset of 60 vendors and 220 banks that loss distributions are heavy-tailed and highly sensitive to how quickly patches are applied. They also develop an early-warning graph neural network (CFC-GNN) that achieves AUROC 0.82 and AUPRC 0.60 in flagging high-risk vendors before cascades occur. The paper concludes that concentration of banking operations among a small number of AI vendors is a first-order financial stability problem and provides supervisors with a quantitative tool for assessing that risk.
- AI policy
- Enterprise
Research
DiSCo: A Distribution-First Steering and Cultural Prior Evaluation Framework for Measuring Cultural Preference Bias in LLMs
Bhuvan Arora, Devesh Saraogi, Sravya Varada et al.
arXiv · 2026-09-09
DiSCo is a new evaluation framework that measures how large language models (LLMs) default to certain cultural preferences and whether those defaults can be changed through prompting. Using a 304-item benchmark spanning 12 cultures, the authors find that UK and US cultures together absorb roughly 35% of all model selections despite representing only 2 of the 12 cultures tested. Critically, prompt-based steering does not fix the bias — it actually widens the gap between high- and low-resource cultures, and injecting explicit cultural facts produces negligible change. These findings show that cultural preference bias in LLMs is deeply embedded and cannot be resolved through prompt-based personalisation alone, raising important concerns about equitable deployment of AI assistants globally.
- AI policy
- Quality assurance
Research
Can AI Agents Deliver Verifiable Network-Wide Outcomes Across Authority Boundaries?
Tianzhu Zhang, Chih-Kai Huang, Meikang Qiu
arXiv · 2026-09-09
This paper addresses a critical challenge in AI-driven network automation: when multiple AI agents with distinct authority scopes coordinate configuration changes across a network, there is no reliable way to confirm that the intended network-wide outcome was actually achieved. The authors introduce EvidenceNet, a runtime assurance layer that collects post-change observations from all required authority scopes, verifies their freshness and provenance, and uses a verifier agent to assess whether the collective evidence satisfies the operator's intended outcome. Experiments on live routing networks demonstrate that checking post-change state reveals successful outcomes that configuration-action records alone cannot confirm, and controlled interventions show EvidenceNet correctly rejects completion when observations are stale, substituted, or sourced from the wrong scope. This matters because it provides a verifiable basis for trusting coordinated AI agent actions in operational networks spanning multiple administrative domains.
- Quality assurance
- Enterprise
Research
Beyond Training: A Feasibility Taxonomy for Inference-Time AI Governance
Samar Ansari
arXiv (Cornell University) · 2026-09-09
This paper argues that current AI governance frameworks focus almost entirely on training compute, but increasingly miss the point as AI capabilities migrate to the inference stage through scaling, agentic use, and on-device deployment. The authors build a taxonomy of twenty inference-time governance mechanisms spanning monitoring, verification, and enforcement, rating each on a four-point readiness scale using evidence from four vendors. While fifteen of the twenty mechanisms have commercial technical substrates already in production, the analysis reveals that robustness against adversaries is highly uneven: no mechanism rates as adequate against a high-capability state-level deployer, and fine-tuning can strip model-internal enforcement components. The paper maps these mechanisms to four governance scenarios and links them to a companion hardware-focused paper via a conditional substitution principle, offering a structured foundation for policymakers seeking to regulate AI at the point of deployment rather than at training time.
- AI policy
- Certifications
Research
Reference-Based Bias Detection in LLMs via Relative Representations of Hidden States
Marek Jeliński, Jan Dubiński, Maciej Chrabaszcz et al.
arXiv · 2026-09-09
This paper proposes a reference-based method for detecting bias in large language models (LLMs) by analyzing internal hidden-state representations rather than relying solely on model outputs. The approach encodes sentences relative to anchor sentences, measuring how target group associations with positive and negative attributes shift after fine-tuning — a metric called Representational Bias Shift (ΔB). Tested across three model families and three benchmarks (WildGuardMix, DecodingTrust, ToxiGen), ΔB correlates with output-level bias change in 15 of 18 settings, achieves ROC AUC between 0.65 and 0.99 for detecting increased bias, and requires 3–50× less compute than output-level benchmarks. This matters for AI quality assurance and policy because it enables faster, cheaper bias auditing of model checkpoints without needing task-specific evaluation data or expensive judge models.
- Quality assurance
- AI policy
Research
MedDeID enables locally governed clinical-text de-identification from real or synthetic training data
Stig Hellemans, Tom Stroobants, Elyne Scheurwegs et al.
arXiv · 2026-09-09
MedDeID is an on-premises framework for de-identifying clinical notes — removing personally identifiable information (PII) — without requiring patient data to leave a hospital. Tested on a 300-note Dutch hospital benchmark, a hospital-trained compact transformer detected 98.9% of identifying text, while a model trained solely on synthetic notes detected 96.1%; on 100 primary-care notes the synthetic model actually outperformed the hospital-trained model in recall (90.3% vs. 87.0%). An English version trained without real patient text also performed strongly on synthetic benchmarks, demonstrating the workflow transfers across languages. The framework matters because it lets healthcare institutions meet privacy requirements while still enabling clinical AI research using locally governed, real or synthetic training data.
- AI policy
- Enterprise
Research
What Makes Adversarial Examples Transfer Across Deepfake Detectors?
Rafael M. Mamede, Pedro C. Neto, Ana F. Sequeira
arXiv · 2026-09-09
This paper investigates why adversarial examples crafted on one deepfake detector successfully fool a different detector — a threat known as black-box transfer attack. Using a controlled benchmark of 60 detectors spanning six backbone architectures, two pretraining regimes, and five training-data configurations, the authors find that transfer success is strongly shaped by how closely the source and target models share architectural or training characteristics. Mean attack success rates range from roughly 7% to 20% for single-source attacks, but a multi-source oracle combining both evaluated attack methods reaches 64.48% success even when excluding exact backbone or training-data matches — demonstrating that standard single-source evaluations can severely understate how vulnerable detectors actually are. The results matter for deepfake detection quality assurance and policy, establishing that source–target compatibility and source-model selection are critical, often overlooked dimensions of credible robustness evaluation.
- Quality assurance
- AI policy
Research
When Does Defendant Statement Matter? A Study of Bias and Persuasion in LLM-Simulated Jurors
Cho-Ying Wu
arXiv · 2026-09-09
This paper examines how LLM-simulated jurors respond to defendant courtroom statements in U.S. criminal trials, introducing JuryBench, a benchmark of controversial criminal cases with defendants of varied backgrounds and statements of differing emotional appeal. Across 20 frontier LLMs generating 432K decisions, the study finds that emotional persuasion can backfire by signaling guilt or inconsistency, that background similarity between juror and defendant is a stronger predictor of verdict severity than other factors, and that juror ideology significantly shapes judgments. The findings reveal that LLM jury simulations reproduce several known human-jury biases, raising both opportunities and risks for using LLMs to model legal reasoning and highlighting the need for careful evaluation before deployment in legal contexts.
- AI policy
- Quality assurance
Research
Scored vs. Generated Readouts in Behavioral Language Models: An Empirical Study of Elicitation Format
Touchapon Kraisingkorn, Krittin Pachtrachai, Wachiravit Modecrua
arXiv · 2026-09-09
This paper compares two ways of extracting predictions from behavioral language models fine-tuned on retail customer data: directly scoring answer-token probabilities versus generating a written rationale first and then reading off the prediction. Across 13 model-domain test cells covering four retail tasks in three markets, scored readouts rank outcomes more accurately in 12 of 13 cases (AUC improvements of 1.5 to 14.5 points), while generated rationales show reduced reliance on dominant predictive features and tend toward stock formulations. A third approach—eliciting an explicit probability before any verdict—improves calibration (Brier score from 0.47 to 0.15) but only for outcome rates represented in training. The authors recommend using generated rationales for explanations while sourcing ranking signals from the scored head, a practical finding relevant to enterprises deploying language models for customer behavior prediction.
- Enterprise
Research
AgentAudit: An Open, Extensible Framework for Full-Lifecycle Trust Evaluation of AI Agents
Shrey Nag, Sachita, Abhishek Kumar Singh et al.
arXiv · 2026-09-09
AgentAudit introduces an open evaluation framework that assesses AI agents across their full execution lifecycle—covering planning, tool selection, tool execution, memory, and reasoning—rather than just task completion or isolated security robustness. The framework attaches to any LLM-based agent without modifying it, reading only recorded execution traces to score ten dimensions and pinpoint the exact stage where failures occur. Evaluating five models (GPT-5, Claude Sonnet 5, Sarvam 105B, Llama 3.3 70B, Gemini 2.5 Flash), the study finds large trust score gaps (95.1 down to 22.6 out of 100) and reveals that some models actively comply with adversarial instructions rather than simply failing them—a distinction pass/fail benchmarks cannot surface. This matters because it provides a richer, auditable basis for determining whether AI agents can be trusted in real-world deployments.
- Quality assurance
- Certifications
Research
Subgroup Membership Inference Audits of Differentially Private Synthetic Text
Yidan Sun, Viktor Schlegel, Srinivasan Nandakumar et al.
arXiv · 2026-09-09
This paper investigates whether differentially private (DP) synthetic text releases adequately protect vulnerable subgroups from membership inference attacks (MIAs). The authors define a subgroup-targeted membership inference game and audit 32 subgroup proxies across four datasets, three generators (DP-SGD fine-tuning, API-based prompting, and activation steering), and five privacy budgets, finding that synthetic releases leak subgroup membership and that prior attacks systematically underestimate this leakage. While DP reduces average leakage at every tested budget, the remaining risk is concentrated—roughly 10% of records carry about 40% of residual leakage—and DP removes more leakage from random records than from high-risk ones, meaning protection is uneven within its worst-case guarantee. The findings matter for policy and quality assurance because they show that aggregate privacy metrics can obscure disproportionate risks to specific subgroups, and that record-level risk cannot be assessed independently of the release mechanism.
- AI policy
- Quality assurance
Research
MUCnoHARM@GermEval Shared Task 2026: Retrieval-based In-Context Learning for Defamatory Offences, and Where It Falls Short
Kristin Gnadt, Maximilian Meidinger, Matthias Aßenmacher
arXiv · 2026-09-09
This paper investigates retrieval-based in-context learning (RetICL) strategies for automatically detecting defamatory offences under German criminal law (§§ 185-187 StGB) in social media posts, as part of the GermEval 2026 Shared Task. The study finds that few-shot prompting outperforms zero-shot approaches, but retrieval-based methods offer only marginal improvements over randomly selected demonstrations and can even underperform a carefully chosen static demonstration set. Model selection proves to be the most influential factor, surpassing the impact of prompting strategy or legal knowledge injection. Because tested models over-predict criminal relevance while still missing 26-57% of truly criminal posts, the authors conclude they are better suited for triage support than autonomous moderation decisions.
- AI policy
- Quality assurance
Research
When Auditors Fabricate: Batch-Size Degradation and Confident Hallucination in LLM Detection of Planted Document Contamination
Karan Parekh, Sanjana Pendyala Ravinder, Sana Mhapsekar et al.
arXiv · 2026-09-09
This paper evaluates how reliably Google Gemini 2.0 Pro can detect deliberately planted errors in a corpus of 150 academic papers (450 injected contaminants of three types: typographical corruption, semantic reversal, and absurd out-of-context insertion). Detection performance collapses at scale — dropping from 50–60% recovery on single or small-batch documents to just 2.8% on large batches — and the failure mode is not abstention but confident fabrication of non-existent errors, including invented phrases like 'telepathic squirrel' and 'quantum-powered toaster.' Critically, the most realistic contaminants (typographical and semantic errors) were detected least reliably (50% each), while absurd insertions were caught most often (75%), meaning the errors most likely to appear in real documents are the ones most likely to be missed. The authors conclude that LLM-based document auditing degrades deceptively rather than gracefully, and recommend bounded batch sizes, direct content injection, and mechanical verification of every reported finding against source text.
- Quality assurance
Research
Chance, Persistent Advantage, and the Generative-AI Era in Open-Source Package Careers
Hazem Ibrahim, Yasir Zaki
arXiv (Cornell University) · 2026-09-09
This study examines whether patterns of career success documented in science and the arts—where timing of a biggest hit is essentially random but individual capacity for large successes is stable—also apply to open-source software contributors. Analyzing 102.2 million career works by 6.15 million GitHub contributors (2015–2025) and measuring impact by downstream package dependencies, the authors find that the timing of a contributor's biggest hit is close to a lottery over their works, a stable personal factor explains only about one-fifth of impact persistence while the rest reflects momentum, and these patterns largely mirror those seen in creative fields. Crucially, the arrival of generative AI tools like ChatGPT produced no detectable shift in this structure—the estimated effect was +0.03 (95% CI [−0.22, +0.23]), indistinguishable from zero after accounting for cohort aging. These findings have implications for how open-platform track records should be interpreted and for realistic expectations about what generative AI will do to individual software careers.
- Workforce
- Enterprise