Search: “safety”

And eventually schools will (if they have any sense) probably withdraw or limit LLMs.

And eventually schools will (if they have any sense) probably withdraw or limit LLMs. 🦔Schools across the US are reversing years of technology-first classroom policies after studies show laptop and sc

safetygary-marcus--x

Safety

Clinical Reasoning AI for Oncology Treatment Planning: A Multi-Specialty Case-Based Evaluation

DGX agent

arXiv:2604.20869v1 Announce Type: cross Abstract: Background: More than 80% of U.S. cancer care is delivered in community settings, where survival remains worse than at academic centers. Clinicians mu

safetyarxiv-cs-ai

24 Apr 2026

FryNet: Dual-Stream Adversarial Fusion for Non-Destructive Frying Oil Oxidation Assessment

arXiv:2604.21321v1 Announce Type: new Abstract: Monitoring frying oil degradation is critical for food safety, yet current practice relies on destructive wet-chemistry assays that provide no spatial i

Seeing Without Eyes: 4D Human-Scene Understanding from Wearable IMUs

arXiv:2604.21926v1 Announce Type: new Abstract: Understanding human activities and their surrounding environments typically relies on visual perception, yet cameras pose persistent challenges in priva

Value-Conflict Diagnostics Reveal Widespread Alignment Faking in Language Models

arXiv:2604.20995v1 Announce Type: new Abstract: Alignment faking, where a model behaves aligned with developer policy when monitored but reverts to its own preferences when unobserved, is a concerning

DAIRE: A lightweight AI model for real-time detection of Controller Area Network attacks in the Internet of Vehicles

arXiv:2604.20771v1 Announce Type: cross Abstract: The Internet of Vehicles (IoV) is advancing modern transportation by improving safety, efficiency, and intelligence. However, the reliance on the Cont

Diagnosing CFG Interpretation in LLMs

arXiv:2604.20811v1 Announce Type: new Abstract: As LLMs are increasingly integrated into agentic systems, they must adhere to dynamically defined, machine-interpretable interfaces. We evaluate LLMs as

Evaluating Assurance Cases as Text-Attributed Graphs for Structure and Provenance Analysis

arXiv:2604.20577v1 Announce Type: cross Abstract: An assurance case is a structured argument document that justifies claims about a system's requirements or properties, which are supported by evidence

Generative Augmentation of Imbalanced Flight Records for Flight Diversion Prediction: A Multi-objective Optimisation Framework

arXiv:2604.20288v1 Announce Type: new Abstract: Flight diversions are rare but high-impact events in aviation, making their reliable prediction vital for both safety and operational efficiency. Howeve

Graph2Counsel: Clinically Grounded Synthetic Counseling Dialogue Generation from Client Psychological Graphs

arXiv:2604.20382v1 Announce Type: new Abstract: Rising demand for mental health support has increased interest in using Large Language Models (LLMs) for counseling. However, adapting LLMs to this high

More people will die from suppressing AI than from the imaginary AI apocalypse. They'll die from restricting safe self-driving cars that are…

More people will die from suppressing AI than from the imaginary AI apocalypse. They'll die from restricting safe self-driving cars that are 90% better drivers than people who kill 1.5 million people

safetyyann-lecun--x

Toward Cooperative Driving in Mixed Traffic: An Adaptive Potential Game-Based Approach with Field Test Verification

arXiv:2604.20231v1 Announce Type: new Abstract: Connected autonomous vehicles (CAVs), which represent a significant advancement in autonomous driving technology, have the potential to greatly increase

AutoAWG: Adverse Weather Generation with Adaptive Multi-Controls for Automotive Videos

arXiv:2604.18993v1 Announce Type: cross Abstract: Perception robustness under adverse weather remains a critical challenge for autonomous driving, with the core bottleneck being the scarcity of real-w

Benchmarking Misuse Mitigation Against Covert Adversaries

arXiv:2506.06414v2 Announce Type: replace-cross Abstract: Existing language model safety evaluations focus on overt attacks and low-stakes tasks. In reality, an attacker can easily subvert existing sa

Ensembling Pruned Attention Heads For Uncertainty-Aware Efficient Transformers

arXiv:2510.18358v2 Announce Type: replace-cross Abstract: Uncertainty quantification (UQ) is essential for deploying deep neural networks in safety-critical settings. Although methods like Deep Ensemb

From Particles to Perils: SVGD-Based Hazardous Scenario Generation for Autonomous Driving Systems Testing

arXiv:2604.18918v1 Announce Type: cross Abstract: Simulation-based testing of autonomous driving systems (ADS) must uncover realistic and diverse failures in dense, heterogeneous traffic. However, exi

Mind2Drive: Predicting Driver Intentions from EEG in Real-world On-Road Driving

arXiv:2604.19368v1 Announce Type: new Abstract: Predicting driver intention from neurophysiological signals offers a promising pathway for enhancing proactive safety in advanced driver assistance syst

REVEAL: Multimodal Vision-Language Alignment of Retinal Morphometry and Clinical Risks for Incident AD and Dementia Prediction

arXiv:2604.18757v1 Announce Type: cross Abstract: The retina provides a unique, noninvasive window into Alzheimer's disease (AD) and dementia, capturing early structural changes through morphometric f

When Can We Trust Deep Neural Networks? Towards Reliable Industrial Deployment with an Interpretability Guide

arXiv:2604.19206v1 Announce Type: new Abstract: The deployment of AI systems in safety-critical domains, such as industrial defect inspection, autonomous driving, and medical diagnosis, is severely ha

Alignment Data Map for Efficient Preference Data Selection and Diagnosis

arXiv:2505.23114v3 Announce Type: replace Abstract: Human preference data is essential for aligning large language models (LLMs) with human values, but collecting such data is often costly and ineffic

Asset Harvester: Extracting 3D Assets from Autonomous Driving Logs for Simulation

arXiv:2604.18468v1 Announce Type: new Abstract: Closed-loop simulation is a core component of autonomous vehicle (AV) development, enabling scalable testing, training, and safety validation before rea

ASTRA: An Automated Framework for Strategy Discovery, Retrieval, and Evolution for Jailbreaking LLMs

arXiv:2511.02356v2 Announce Type: replace-cross Abstract: Despite extensive safety alignment, Large Language Models (LLMs) remain vulnerable to jailbreak attacks. However, existing methods generally l

Autonomous Vehicle Collision Avoidance With Racing Parameterized Deep Reinforcement Learning

arXiv:2604.16702v1 Announce Type: new Abstract: Road traffic accidents are a leading cause of fatalities worldwide. In the US, human error causes 94% of crashes, resulting in excess of 7,000 pedestria

Characterizing Model-Native Skills

arXiv:2604.17614v1 Announce Type: cross Abstract: Skills are a natural unit for describing what a language model can do and how its behavior can be changed. However, existing characterizations rely on

Cognitive Policy-Driven LLM for Diagnosis and Intervention of Cognitive Distortions in Emotional Support Conversation

arXiv:2604.17178v1 Announce Type: new Abstract: Emotional Support Conversation (ESC) plays a critical role in mental health assistance by providing accessible psychological support in real-world appli

Detecting Alarming Student Verbal Responses using Text and Audio Classifier

arXiv:2604.16717v1 Announce Type: new Abstract: This paper addresses a critical safety gap in the use Automated Verbal Response Scoring (AVRS). We present a novel hybrid framework for troubled student

Local Ai

Different Paths to Harmful Compliance: Behavioral Side Effects and Mechanistic Divergence Across LLM Jailbreaks

arXiv:2604.18510v1 Announce Type: cross Abstract: Open-weight language models can be rendered unsafe through several distinct interventions, but the resulting models may differ substantially in capabi

local-aiarxiv-cs-cl

Efficient Diffusion Models under Nonconvex Equality and Inequality constraints via Landing

arXiv:2604.17838v1 Announce Type: new Abstract: Generative modeling within constrained sets is essential for scientific and engineering applications involving physical, geometric, or safety requiremen

Erasing Thousands of Concepts: Towards Scalable and Practical Concept Erasure for Text-to-Image Diffusion Models

arXiv:2604.16481v1 Announce Type: new Abstract: Large-scale text-to-image (T2I) diffusion models deliver remarkable visual fidelity but pose safety risks due to their capacity to reproduce undesirable

HAVEN: Hierarchical Adversary-aware Visibility-Enabled Navigation with Cover Utilization using Deep Transformer Q-Networks

arXiv:2512.00592v2 Announce Type: replace Abstract: Autonomous navigation in partially observable environments requires agents to reason beyond immediate sensor input, exploit occlusion, and ensure sa

Online Conformal Prediction with Adversarial Semi-bandit Feedback via Regret Minimization

arXiv:2604.17984v1 Announce Type: new Abstract: Uncertainty quantification is crucial in safety-critical systems, where decisions must be made under uncertainty. In particular, we consider the problem

The Impact of Off-Policy Training Data on Probe Generalisation

arXiv:2511.17408v4 Announce Type: replace-cross Abstract: Probing has emerged as a promising method for monitoring large language models (LLMs), enabling cheap inference-time detection of concerning b

They just indicted the Southern Poverty Law Center on 11 counts of fraud. Wow! The SPLC worked with Facebook, Old Twitter and Youtube on the…

They just indicted the Southern Poverty Law Center on 11 counts of fraud. Wow! The SPLC worked with Facebook, Old Twitter and Youtube on their safety teams, sometimes directly advising/flagging what c

safetyelon-musk--x

AI Agents and Hard Choices

arXiv:2504.15304v2 Announce Type: replace Abstract: Can AI agents deal with hard choices -- cases where options are incommensurable because multiple objectives are pursued simultaneously? Adopting a t

Jailbreak Scaling Laws for Large Language Models: Polynomial-Exponential Crossover

arXiv:2603.11331v2 Announce Type: replace-cross Abstract: Adversarial attacks can reliably steer safety-aligned large language models toward unsafe behavior. Empirically, we find that strong adversari

Safe and Energy-Aware Multi-Robot Density Control via PDE-Constrained Optimization for Long-Duration Autonomy

arXiv:2604.15524v1 Announce Type: cross Abstract: This paper presents a novel density control framework for multi-robot systems with spatial safety and energy sustainability guarantees. Stochastic rob

Safe Deep Reinforcement Learning for Building Heating Control and Demand-side Flexibility

arXiv:2604.16033v1 Announce Type: cross Abstract: Buildings account for approximately 40% of global energy consumption, and with the growing share of intermittent renewable energy sources, enabling de

Dario is not as different from Sam as people seem to think.

Gary Marcus argues that Dario Amodei (Anthropic CEO) and Sam Altman (OpenAI CEO) share more similarities in their approaches to AI development and safety than commonly perceived, despite their public

safetygary-marcus--x

18 Apr 2026

Context Over Content: Exposing Evaluation Faking in Automated Judges

arXiv:2604.15224v1 Announce Type: cross Abstract: The extit{LLM-as-a-judge} paradigm has become the operational backbone of automated AI evaluation pipelines, yet rests on an unverified assumption: th

[Emerging Ideas] Artificial Tripartite Intelligence: A Bio-Inspired, Sensor-First Architecture for Physical AI

arXiv:2604.13959v1 Announce Type: new Abstract: As AI moves from data centers to robots and wearables, scaling ever-larger models becomes insufficient. Physical AI operates under tight latency, energy

From Risk to Rescue: An Agentic Survival Analysis Framework for Liquidation Prevention

arXiv:2604.14583v1 Announce Type: new Abstract: Decentralized Finance (DeFi) lending protocols like Aave v3 rely on over-collateralization to secure loans, yet users frequently face liquidation due to

Integrating Object Detection, LiDAR-Enhanced Depth Estimation, and Segmentation Models for Railway Environments

arXiv:2604.14781v1 Announce Type: new Abstract: Obstacle detection in railway environments is crucial for ensuring safety. However, very few studies address the problem using a complete, modular, and

Optimistic Policy Learning under Pessimistic Adversaries with Regret and Violation Guarantees

arXiv:2604.14243v1 Announce Type: new Abstract: Real-world decision-making systems operate in environments where state transitions depend not only on the agent's actions, but also on extbf{exogenous f

The Autocorrelation Blind Spot: Why 42% of Turn-Level Findings in LLM Conversation Analysis May Be Spurious

arXiv:2604.14414v1 Announce Type: new Abstract: Turn-level metrics are widely used to evaluate properties of multi-turn human-LLM conversations, from safety and sycophancy to dialogue quality. However

C^2T: Captioning-Structure and LLM-Aligned Common-Sense Reward Learning for Traffic--Vehicle Coordination

arXiv:2604.13098v1 Announce Type: cross Abstract: State-of-the-art (SOTA) urban traffic control increasingly employs Multi-Agent Reinforcement Learning (MARL) to coordinate Traffic Light Controllers (

From Where Words Come: Efficient Regularization of Code Tokenizers Through Source Attribution

arXiv:2604.14053v1 Announce Type: new Abstract: Efficiency and safety of Large Language Models (LLMs), among other factors, rely on the quality of tokenization. A good tokenizer not only improves infe

Learning Probabilistic Responsibility Allocations for Multi-Agent Interactions

arXiv:2604.13128v1 Announce Type: cross Abstract: Human behavior in interactive settings is shaped not only by individual objectives but also by shared constraints with others, such as safety. Underst

Robust Energy-Aware Routing for Air-Ground Cooperative Multi-UAV Delivery in Wind-Uncertain Environments

arXiv:2604.13441v1 Announce Type: new Abstract: Ensuring energy feasibility under wind uncertainty is critical for the safety and reliability of UAV delivery missions. In realistic truck-drone logisti