Large language models (LLMs) increasingly support science, but they can also convert hazardous scientific knowledge into actionable misuse guidance. Existing benchmarks often rely on templated queries disconnected from real-world hazards, and employ LLM-as-a-Judge paradigms without domain grounding.
arXiv:2512. 15783v3 Announce Type: replace-cross Abstract: This paper proposes a measurement standardisation framework that compresses expert-AI interactions into structured, comparable fields for prospective risk detection in deployed AI systems, without access to model internals.
arXiv:2607. 18665v1 Announce Type: new Abstract: Large language models (LLMs) increasingly support science, but they can also convert hazardous scientific knowledge into actionable misuse guidance.
Claims about AI safety reach audiences well beyond the AI community, yet many rely on opaque evidence or static assessments, when supporting evidence is accessible at all. We present the Systemic Risk...
The paper introduces the Systemic Risk Index, an open pipeline and dashboard that aggregates evidence from 19 public AI benchmarks into four systemic‑risk categories defined by the EU GPAI Code of Practice. It evaluates 18 models using harm‑preserving perturbations and simulated deployment contexts, offering users the ability to switch between average and worst‑case aggregation and to trace each risk rating back to its benchmark evidence. The study finds that worst‑case scores can be 14 to 37 points lower than average scores, and that LLM judges agree with human graders at a level comparable to human‑human agreement.
arXiv:2607. 19292v1 Announce Type: cross Abstract: Current AI safety discourse still focuses disproportionately on visible failures, including obvious harms, dramatic misuse, and hypothetical catastrophic scenarios.
The paper introduces ARGUS, a language‑model pipeline that audits evidence for identification assumptions in difference‑in‑differences studies of climate policy. ARGUS evaluates reported evidence against an eleven‑dimension rubric, abstaining when evidence cannot be retrieved. In tests, ARGUS detects 73% of injected flaws versus 18% for a keyword approach, abstains on about 40% of assessments in 26 economics papers, and often assigns higher risk than human labels in a five‑paper pilot.
arXiv:2606. 30953v1 Announce Type: new Abstract: We introduce the Neuro-Bayesian-Symbolic Residual Attention Shallow Network (NBS-RASN), a hybrid neural architecture for explainable cybersecurity risk assessment in open-source ecosystems.
The study evaluates 684 U.S. federal AI governance documents for how they address 14 sectors and 24 AI risks, measuring both breadth and depth of coverage. It finds that risks related to robustness, system security, and governance are more frequently and substantively discussed than socioeconomic, environmental, and emerging risks, and that sectors such as public administration, national security, information, and scientific services receive higher coverage than finance and healthcare. By comparing these coverage patterns with expert vulnerability assessments, the authors identify potential gaps in AI governance that could inform future policy and industry decisions.
arXiv:2607. 23983v1 Announce Type: cross Abstract: Operational flood forecasting depends on tacit forecaster expertise that is difficult to formalize, audit, and transfer.
arXiv:2602. 05056v2 Announce Type: replace-cross Abstract: Online scams increasingly leverage fluent and context-aware social engineering strategies, creating growing demand for AI systems that explain why a message may be risky.
The paper introduces a deterministic AI security risk assessment framework that transforms diverse engineering artefacts into a standardized Control ID taxonomy scored on a four‑level ordinal scale. It compiles technique‑level predicates from a fixed MITRE ATLAS snapshot, linking each control to mitigation and producing traceable feasibility and impact outputs. The framework is formally verified for boundedness, totality, consistency, and monotonicity, and is evaluated on five open‑source AI projects, showing that strengthened controls lower feasibility scores while residual risks persist when core controls are missing.