Search: “safety-filtering”

Smart search ranks by meaning as well as keywords (one row per story, last 45 days).

Safety for Whom? Boundary-Aware Self-Distillation for Controlled LLM Safety Refusal

A self-distillation safety framework tunes narrow-boundary refusals in Qwen3-8B, raising target-domain refusal to 84.75% while cutting over-refusal from 15.20% to 5.20%.

The paper formulates narrow-boundary safety, where deployments need refusals within specific topics rather than whole subjects, and proposes an offline self-generated framework with controlled topic generation, escalating retries, and harmful-benign boundary pairs. On political persuasion with Qwen3-8B, the method raised target-domain refusal from 9.47% to 84.75% and cut the mean unsafe-response rate across three broader benchmarks from 26.26% to 0.14%. Verified target-model responses reduced over-refusal from 15.20% to 5.20%, and boundary-pair data cut comply-side over-refusal on held-out pairs from 32.94% to 4.16%. Results show data composition controls the safety-usability trade-off and alignment should be evaluated on both sides of the refusal boundary.

Hugging Face daily papers·Sep 2, 20:00 UTC · 14d agoAI safety & security1

32

Safe Meta-Reinforcement Learning via Information Space Reachability

Safe meta-RL framework reasons about safety in information space, learning a safety value function used for safety filtering and constrained policy optimization.

The paper proposes safe meta-RL that reasons about safety in information space, capturing both physical state and the agent's belief over the underlying task. A safety value function measures the probability of avoiding unsafe regions indefinitely and satisfies a self-consistency condition and Bellman equation, making it learnable via meta-RL. The resulting algorithm uses the learned function for safety filtering and constrained policy optimization, with effectiveness demonstrated on meta-RL benchmarks.

arXiv cs.AI / cs.LG / cs.CL·Sep 14, 17:27 UTC · 2d agoAI research

25

Perturbation Probing: A New Diagnostic for the Fragility of LLM Safety

Unit 42 research shows LLM safety refusals concentrate in a thin neural layer, motivating external, multi-layered AI security controls.

Palo Alto Networks Unit 42 introduces Perturbation Probing, a diagnostic technique for measuring the fragility of LLM safety mechanisms. The research finds that safety refusal behavior is localized within a thin neural layer, implying small perturbations can undermine built-in refusals. The authors argue this motivates external, multi-layered security defenses on top of model-internal safety training.

Palo Alto Unit 42·Aug 28, 22:00 UTC · 19d agoAI safety & security

45

Recorded Future Launches AI Alert Filtering

Recorded Future launched AI Alert Filtering to automatically triage threat intelligence alerts by relevance, letting analysts prioritize faster while retaining control.

Recorded Future announced general availability of AI Alert Filtering, powered by Recorded Future AI. The feature automates the first pass of filtering alerts by relevance so analysts can prioritize faster while keeping human control. It applies LLM-based automation to threat intelligence alert triage workflows.

Recorded Future·Aug 26, 00:00 UTC · 22d agoTools

25

Operational Resilience: IT Security Risks with Reduced Staffing | Huntress

Huntress blog advises security teams on managing change, risk, and incident response during reduced-staffing holiday periods.

The article discusses how holiday-period staffing reductions change organizational risk profiles around change management, monitoring, and incident response capability. It argues against blanket change freezes when critical vulnerabilities with high exploitation probability demand patching, and stresses retaining decision-making authority, escalation paths, and recovery knowledge. It concludes by promoting Huntress Managed Response, which lets the Huntress SOC take predefined containment actions on confirmed threats without customer intervention.

Huntress·Sep 15, 14:00 UTC · 1d agoIndustry

12

Edge Case launches Guardian, an AI platform for tracking risk across autonomous systems

Edge Case launched Guardian, an AI platform creating a Digital Safety Twin to continuously track risk across autonomous and defense systems.

Edge Case launched Guardian, an AI-driven platform that fuses safety analysis, engineering data, and operational signals into a continuous Digital Safety Twin. It targets advanced autonomous platforms in commercial and defense markets, with six customer agreements signed at launch. The platform tracks hazards, mitigations, and source-linked evidence across the development lifecycle, shifting safety from one-time review to continuous capability.

Help Net Security·Sep 2, 00:00 UTC · 15d agoAI tools & infra

18

Inside AI-Powered WAF Detections: Architecture and Safety Controls

Akamai published a technical breakdown of the architecture and safety controls behind its AI-powered WAF detections.

Akamai published a technical blog post examining the architecture of its AI-powered web application firewall detections. Per the title, the piece covers the design and safety controls applied to machine-learning-based WAF detections.

Akamai Blog·Sep 8, 12:00 UTC · 9d agoTools

15

TIER: Threat Implicitness Benchmark for Evaluating LLM Safety Behaviors

TIER benchmark shows LLM safety behaviors shift gradually across threat implicitness levels, with jailbreaks exposing the largest robustness gaps.

The TIER benchmark evaluates LLM safety behaviors across four risk domains and four threat levels, from explicit harmful requests to sophisticated jailbreaks, using a six-label behavior scale and two independent LLM judges. Experiments on six open-weight LLMs show safety behaviors evolve gradually across threat levels rather than flipping from refusal to compliance. Models with similar Attack Success Rates can exhibit distinct response distributions, arguing for behavior-aware safety evaluation.

arXiv cs.CR·Sep 4, 13:13 UTC · 13d agoAI safety & security

35

SAFIRE: Safety-Critical Benchmark for Fine-grained Fire and Smoke Understanding in Multimodal LLMs

SAFIRE, an 83K-image fire and smoke benchmark, shows open-source multimodal LLMs average only 61.9% accuracy on safety-critical fire reasoning.

SAFIRE is a large-scale benchmark for fire-smoke understanding in multimodal LLMs with 83K captioned images across 20 scenarios and 193K multiple-choice VQA questions spanning 10 evaluation dimensions from perception to higher-order reasoning. Annotations were built via a GPT-5.4-assisted multi-stage pipeline with MLLM majority voting. Ten open-source MLLMs (8B-38B) average 61.9% accuracy, exposing major gaps in safety-critical reasoning. Adapting vision encoders on 7% of the domain data raises fire-scene classification from 20.1% to 64.5%.

arXiv cs.AI / cs.LG / cs.CL·Sep 7, 17:59 UTC · 9d agoAI research

33

Big Tech’s AI safety rift signals disruption and disparity for enterprises

Diverging AI safety stances among major labs will make frontier model access less predictable, pushing enterprises toward routing layers and independent validation.

A public rift among leading AI labs over safety approaches - Meta's Zuckerberg backing neutral evaluators, Dario Amodei urging a slower pace, and Sam Altman calling for collaboration on standards - is creating operational challenges for enterprise IT. Analysts from Gartner and others say divergent vendor release schedules, access tiers, and regional restrictions will make frontier model access less predictable, effectively treating frontier AI as a managed supply with pricing premiums. Recommendations include routing layers between applications and providers, contractual deprecation terms, and independent validation of models before production use.

CSO Online·Sep 16, 16:19 UTC · 20h agoAI industry

32

The Model Proposes, the Code Disposes: A Pre-Registered Ablation of a Verifier-and-Acceptance Stage in an LLM-Orchestrated Offensive-Security Agent

Pre-registered ablation finds a model verifier stage in an LLM offensive-security agent suppresses findings; removing it eliminated suppression with precision tradeoff.

The paper evaluates a verifier-and-acceptance stage in an LLM-orchestrated offensive-security agent via a pre-registered 20-run confirmatory ablation and a 2x2 factorial study with 40 runs on vulnerable lab targets. Removing the stage eliminated pre-report suppression (median 2 vs 0 findings, p = 0.00003) but reduced model-blinded shipped precision (0.471 vs 0.353, p = 0.0087). Suppression was attributed to the model verifier rather than deterministic acceptance rules, and an instrumented canary recorded zero external contacts in all 60 runs. The full design retained 93.8% of model-adjudicated true candidates but failed its pre-registered non-inferiority floor of 0.90.

arXiv cs.CR·Sep 14, 17:13 UTC · 2d agoResearch

38

Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models

Study shows training LLMs on refusal rationales instead of boilerplate refusal statements reduces false refusals while maintaining safety performance.

The paper decomposes safety-tuning responses into a boilerplate refusal statement and an explanatory rationale, finding that refusal statements push models to rely on superficial cues and misjudge benign queries as harmful. Training solely on rationales reduces false refusals while maintaining comparable safety performance, and the benefits carry over to in-context learning configurations and remain compatible with inference-time mitigations. The results argue for precisely curated, fine-grained safety supervision datasets when aligning LLMs.

Hugging Face daily papers·Sep 3, 20:00 UTC · 13d agoAI safety & security1

44

Quantifying IIoT Sensor Node Criticality by Fusing its Data Criticality and Security Vulnerability

Researchers propose a Dempster–Shafer framework fusing IIoT sensor data criticality with CVSS 4.0/3.1 vulnerability scores to rank node criticality.

The paper introduces a framework that evaluates Industrial IoT sensor node criticality by fusing data criticality and cybersecurity vulnerability scores using Dempster–Shafer (D-S) theory. It was validated on a dataset from red wine production and is claimed to generalize to other industrial settings with minimal modification. Results show criticality rankings derived from CVSS 4.0 scores differ significantly from those derived from CVSS 3.1, underscoring how vulnerability scoring methodology affects security prioritization.

arXiv cs.CR·Sep 9, 07:03 UTC · 8d agoResearch

15

CTEM Is Not About the Stages. It’s About the Outcome.

Horizon3 argues CTEM programs should measure continuously reduced exposure rather than mapping technologies to Gartner's five stages.

Horizon3 contends that Continuous Threat Exposure Management should be judged by one outcome: continuously reducing attacker-reachable exposure, not by mapping a technology to each of Gartner's five stages. The post argues validation and verification, not visibility or closed tickets, provide evidence that attack paths are actually broken. It describes a Discover, Validate, Prioritize, Remediate, Verify, Repeat motion as its operationalization of CTEM.

Horizon3.ai·Sep 2, 13:11 UTC · 15d agoIndustry

12

OpenAI previews privacy-focused system for detecting AI misuse

OpenAI previews Private Safety Processing, detecting cross-interaction AI misuse without exposing customer prompt content, rolling out in September.

OpenAI previewed Private Safety Processing for API and enterprise customers, a system that analyzes patterns across related interactions to flag potential misuse while restricting OpenAI personnel from accessing underlying content. It builds on Zero Data Retention (ZDR) deployments, keeping content on customer-controlled infrastructure or a planned option using customer-controlled encryption keys, with prompts and responses not retained after processing. A technical white paper and rollout are planned for September; images flagged as potential CSAM are the retention exception.

Help Net Security·Aug 20, 00:00 UTC · 28d agoAI safety & security

40

When the prompt becomes the payload: A practical pen-testing guide for GenAI, LLM and RAG applications

CSO Online publishes a practical penetration-testing guide for GenAI, LLM, and RAG applications, covering prompt injection, retrieval poisoning, and tenant isolation testing.

The guide frames LLM applications as attack graphs spanning prompts, retrieval layers, vector stores, tools, identities, and downstream APIs, arguing that conventional web testing misses instruction-vs-data channel risks. It builds on OWASP prompt injection guidance (direct vs. indirect injection) and NIST's 2025 adversarial machine-learning taxonomy, noting that RAG and fine-tuning do not remove injection risk. Recommended practices include documenting trust transitions across components, using canaries and synthetic records to avoid test side effects, running multi-turn and obfuscated injection campaigns, and verifying chains from poisoned documents to observable state changes. It also details testing RAG pipelines via controlled document poisoning across metadata, OCR layers, and code comments, plus cross-tenant isolation checks on retrieved document IDs.

CSO Online·Sep 9, 10:00 UTC · 8d agoAI safety & security1

28

The AI policy window is open. We need to act.

OpenAI calls for mandatory national AI safety regulation and backs four California AI safety bills as capabilities accelerate.

OpenAI argues the rapid pace of AI progress, including signs of AI-accelerated research, requires urgent policy action through mandatory, capability-based national regulation. The company endorses four California bills (SB 813, AB 1405, SB 1119, AB 1864) covering independent safety assessments, AI auditor standards, youth protections, and safeguards against AI-enabled biological threats. It also commits to industry-led frontier standards, international coordination, and strengthening internal safeguards such as universal trajectory monitoring and mandatory alignment-evaluation gates for its Astra model. The post references chief scientist Jakub Pachocki's warning about recursive self-improvement and Greg Brockman's "defenders window" concept.

OpenAI News·Sep 9, 13:00 UTC · 8d agoAI policy

72

AI-Driven Threat Intelligence for Gulf Enterprises: Why Detection Speed Is Now a Regulatory Requirement

Cyble argues GCC regulators' 6-72 hour breach-notification deadlines make AI-powered detection essential for Gulf enterprises.

Cyble's blog highlights that the UAE Information Assurance Standard v2 requires incident notification within 6 hours of detection, while Saudi Arabia's SAMA cybersecurity framework and NCA Essential Cybersecurity Controls converge on 72-hour reporting. It argues that compliance clocks start at detection, not response, and that IAS v2 mandates 24/7 monitoring with defined SLAs for Tier 1 critical infrastructure entities. The piece promotes Cyble Vision's AI-powered threat intelligence for continuous exposure monitoring and audit-ready detection logs.

Cyble·Sep 4, 13:23 UTC · 12d agoIndustry

18

MIT creates method to force AI to comply with safety rules

MIT researchers published HardFlow, a method enforcing hard safety constraints on flow-matching generative models' final outputs without retraining.

MIT researchers led by Zeyang Li and Navid Azizan developed HardFlow, a trajectory-optimization method that enforces strict, non-negotiable constraints on flow-matching generative models by checking rule satisfaction only at the final generation step. Published in IEEE TPAMI, it outperformed six rival projection and guidance methods on four simulated benchmarks including D3IL robotic manipulation, Maze2D, physical process control, and image editing. All results are simulation-only, with no independent reproduction yet reported.

Hacker News · security·Sep 14, 16:18 UTC · 2d agoAI researchHN 24↑ · 25 comments

25

Cisco Talos Intelligence blog

Cisco Talos argues restrictive frontier AI models impose a 'safety penalty' that hampers security teams' real-time incident response.

A featured Cisco Talos post contends that increasingly restrictive frontier AI models create a 'safety penalty' slowing defenders' real-time incident response, and advocates 'operational sovereignty' so defensive AI keeps pace with unconstrained adversaries. The blog index also promotes a Beers with Talos podcast episode on how threat intelligence is gathered. The page is a blog listing rather than a single news story.

Cisco Talos·Sep 10, 01:33 UTC · 7d agoAI safety & security

28

The safety penalty: Reclaiming operational sovereignty in the age of AI

Cisco Talos argues restrictive frontier AI models impose a 'safety penalty' on security teams, urging operational sovereignty for defensive AI in incident response.

Cisco Talos published commentary arguing that increasingly restrictive frontier AI models create a 'safety penalty' that slows real-time incident response. It recommends organizations pursue operational sovereignty so defensive AI can keep pace with unconstrained adversaries.

Cisco Talos·Aug 25, 10:00 UTC · 23d agoIndustry

20

Building a risk-based vulnerability management program that scales

Asimily CEO Shankar Somasundaram outlines a risk-based vulnerability management approach using inventory, attack paths, KEV and EPSS data.

In a Help Net Security video, Asimily CEO Shankar Somasundaram argues patching everything is infeasible as AI-driven attacks inflate vulnerability counts, with one customer finding a thousand unknowns for each known one. He recommends building a full inventory of devices, applications, and data flows, mapping attack paths for reachability, and prioritizing with KEV, EPSS, and business impact. Mitigations include patching, virtual patching via NACs and firewalls, segmentation, and configuration snapshots to detect drift.

Help Net Security·Aug 24, 04:21 UTC · 24d agoIndustry

15

How to build an exposure management program the business trusts: Lessons from Tenable’s CSO

Tenable's CSO describes an AI-driven exposure management program that consolidates tool sprawl and translates cyber risk for boards.

A Tenable blog post shares lessons from CSO Robert Huber on moving to an AI-driven exposure management program. It argues tool sprawl and data silos hinder holistic risk assessment and that exposure management unifies attack-surface data into business-level metrics for the C-suite.

Tenable Blog·Aug 27, 14:30 UTC · 20d agoIndustry

16

An Empirical Analysis of ReDoS Vulnerabilities and ReDoS Detection Tools

Study of NVD data finds ReDoS vulnerabilities growing more prevalent and more likely to be exploited, while five detection tools disagree substantially.

The study compares five publicly available ReDoS detection tools and one regex correction tool across three datasets. An empirical analysis of all ReDoS vulnerabilities reported to the NVD finds they are becoming more prevalent and are much more likely to be exploited than non-ReDoS weaknesses. The detection tools exhibited substantial disagreement on whether a given regex is vulnerable.

arXiv cs.CR·Sep 9, 15:09 UTC · 7d agoResearch

35

What researchers learned about building an LLM security workflow

Oslo and FFI researchers show structured agentic workflows lift LLM alert-triage accuracy from 0% to about 93% on malicious cases.

Researchers at the University of Oslo and the Norwegian Defence Research Establishment tested GPT-5-mini, Claude 3 Haiku, Qwen3:30B, and Gemma 3:27B on alerts from the AIT Log Data Set V1.1; given only alert descriptions and log summaries, all four models correctly flagged zero percent of true-positive cases involving reconnaissance, brute-force logins, and initial access. Wrapping the same models in a workflow with constrained SQL queries over Suricata logs, an evidence summarizer, and a verdict stage with revision loops raised malicious-case accuracy to an average of 93 percent, with GPT-5-mini identifying every malicious case across 100 runs. The authors flag it as a proof-of-concept on one synthetic scenario and note models skewed conservative on benign alerts, with GPT-5-mini marking every benign case uncertain.

Help Net Security·Aug 24, 04:21 UTC · 24d agoAI research1

55

Closing the Gap Between Detection and Protection with AI-Assisted Custom Rules

Akamai describes using AI-assisted custom rules to close the gap between threat detection and active protection in security operations.

Akamai published a blog post on AI-assisted custom rules intended to close the gap between detecting threats and enforcing protections. The post appears to be a vendor capability discussion for security operations teams. No article text was available beyond the title, so further technical details are limited.

Akamai Blog·Aug 24, 06:00 UTC · 24d agoTools2

22

How Fragile Is Safety Alignment at Frontier Scale? A Single-Direction Attack on a 320B MoE

Researchers show directional ablation breaks refusal in GLM-5.3-Flash, a 320B-parameter MoE, cutting refusal by 41–89 points across seven benchmarks.

The study extends directional ablation, a white-box attack that removes an aligned LLM's refusal behavior, from dense models up to ~70B parameters to GLM-5.3-Flash, a 320B-parameter mixture-of-experts model with 288 routed experts, four-wide hyper-connection residual, and block-FP8 quantization. Editing attention, dense, and routed-expert writers jointly removes 0.776 of refusal, with 74% of the effect existing only under the joint intervention; the conventional module-name-based recipe reaches only 0.066 and fails silently on MoE architectures. The attack yields 41–89 percentage-point reductions in refusal across seven harmful benchmarks with no detected capability change, and a category-concentrated refusal residue survives all edits at ranks 1 to 12.

arXiv cs.CR·Sep 9, 06:48 UTC · 8d agoAI safety & security

45

Best Practices for Good Endpoint Hardening | Huntress

Huntress outlines endpoint hardening best practices, citing exposed RDP, RMM tool abuse, and ClickFix social engineering as common SMB intrusion paths.

Huntress published defensive guidance on endpoint hardening for small and mid-sized businesses, drawing on observations from its SOC. The post describes common intrusion vectors: internet-exposed RDP brute-forced by automated scanners, phishing emails delivering attacker-controlled remote monitoring and management (RMM) tools, with Huntress reporting a 277% spike in RMM abuse in 2025, and ClickFix attacks using fake CAPTCHA pages that trick users into running malicious commands. Recommended controls include scanning for exposed RDP, SSH, and VPN interfaces, removing unneeded local admin rights, enabling Windows Defender tamper protection, disabling SMBv1, and standardizing on one approved remote access tool, guided by CIS and NIST frameworks.

Huntress·Sep 10, 15:00 UTC · 6d agoAdvisory

20

Your Critical Vulnerabilities Might Not Be Your Biggest Risk

Autonomous penetration testing advocates prioritize exploitable attack paths over raw vulnerability severity for continuous security validation.

The article argues that scanner severity scores lack context: a critical flaw behind strong segmentation may be low priority, while a medium flaw on internet-facing systems can provide a foothold chained toward sensitive data. It positions autonomous penetration testing and attack path validation as the execution layer for continuous security validation, replacing point-in-time assessments. The piece is vendor-authored thought leadership rather than incident or vulnerability news.

The Hacker News·Sep 11, 11:30 UTC · 6d agoIndustry1

18

What Breaks Under Pruning in Smart Homes, and When? Evaluating LLM Degradation Across Architectures and Task Complexity

Pruning study across four LLM architectures finds dense models degrade sharply on smart-home tool calling while MoE models tolerate far more.

Researchers systematically study pruning-induced degradation in smart-home tool calling across four LLMs spanning dense Transformer, dense hybrid, and mixture-of-experts architectures, combining depth, width, hybrid, and expert pruning methods, and evaluate over 19,500 instances from three datasets after post-pruning supervised fine-tuning. Dense models show narrow safe pruning regions followed by sharp degradation, while MoE models tolerate substantially more pruning. Pruning degrades grounded specificity (operation, device, argument, value) before schema-level intent, and aggressive dense pruning can induce systematic over-refusal.

arXiv cs.AI / cs.LG / cs.CL·Sep 15, 17:51 UTC · 1d agoAI research

20