Search: “safety”

Open Challenges in Multi-Agent Security: Towards Secure Systems of Interacting AI Agents

arXiv:2505.02077v2 Announce Type: replace-cross Abstract: AI agents are beginning to interact with each other directly and across internet platforms and physical environments, creating security challe

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Open Problems in Frontier AI Risk Management

DGX agent

arXiv:2604.25982v1 Announce Type: cross Abstract: Frontier AI both amplifies existing risks and introduces qualitatively novel challenges. Not only is there a notable lack of stable scientific consens

safetyarxiv-cs-ai

30 Apr 2026

Roblox reports Q1 bookings up 43% YoY to 1.7B, vs. 1.73B est., and DAUs up 35% to 132M, below analysts' estimates of 143.8M; RBLX drops 16%+ after hours (Cecilia D'Anastasio/Bloomberg)

Cecilia D'Anastasio / Bloomberg: Roblox reports Q1 bookings up 43% YoY to 1.7B, vs. 1.73B est., and DAUs up 35% to 132M, below analysts' estimates of 143.8M; RBLX drops 16%+ after hours — Roblox Corp.

safetytechmeme

SAGE: A Strategy-Aware Graph-Enhanced Generation Framework For Online Counseling

arXiv:2604.26630v1 Announce Type: new Abstract: Effective mental health counseling is a complex, theory-driven process requiring the simultaneous integration of psychological frameworks, real-time dis

safetyarxiv-cs-cl

Sparsity as a Key: Unlocking New Insights from Latent Structures for Out-of-Distribution Detection

arXiv:2604.26409v1 Announce Type: new Abstract: Sparse Autoencoders (SAEs) have demonstrated significant success in interpreting Large Language Models (LLMs) by decomposing dense representations into

Tatemae: Detecting Alignment Faking via Tool Selection in LLMs

arXiv:2604.26511v1 Announce Type: cross Abstract: Alignment faking (AF) occurs when an LLM strategically complies with training objectives to avoid value modification, reverting to prior preferences o

Model Releases

Value-Guided Iterative Refinement and the DIQ-H Benchmark for Evaluating VLM Robustness

model-releasesarxiv-cs-ai

arXiv:2512.03992v2 Announce Type: replace-cross Abstract: Vision-Language Models (VLMs) are essential for embodied AI and safety-critical applications, such as robotics and autonomous systems. However

Knowledge Distillation Must Account for What It Loses

arXiv:2604.25110v1 Announce Type: new Abstract: This position paper argues that knowledge distillation must account for what it loses: student models should be judged not only by retained task scores,

// Latent Agents // Multi-agent debate makes models reason better. It also burns tokens generating long transcripts before any answer comes …

// Latent Agents // Multi-agent debate makes models reason better. It also burns tokens generating long transcripts before any answer comes out. This new research distills the entire debate into a sin

safetydair-ai--x

Reinforcement Learning for Testing Interdependent Requirements in Autonomous Vehicles: An Empirical Study

arXiv:2502.15792v2 Announce Type: replace-cross Abstract: Autonomous vehicles (AVs) make driving decisions without humans, making dependability assurance critical. Scenario-based testing is widely use

VISION-SLS: Safe Perception-Based Control from Learned Visual Representations via System Level Synthesis

arXiv:2604.24894v1 Announce Type: cross Abstract: We propose VISION-SLS, a method for nonlinear output-feedback control from high-resolution RGB images which provides robust constraint satisfaction gu

Betting for Sim-to-Real Performance Evaluation

arXiv:2604.24018v1 Announce Type: new Abstract: This paper studies the problem of robot performance evaluation, focusing on how to obtain accurate and efficient estimates of real-world behavior under

safetyarxiv-cs-ro

Beyond Cross-Modal Alignment: Measuring and Leveraging Modality Gap in Vision-Language Models

arXiv:2502.14888v4 Announce Type: replace-cross Abstract: The success of vision-language models is primarily attributed to effective alignment across modalities such as vision and language. However, m

BMD-45: A Large-Scale CCTV Vehicle Detection Dataset for Urban Traffic in Developing Cities

arXiv:2604.24419v1 Announce Type: new Abstract: Robust vehicle detection from fixed CCTV cameras is critical for Intelligent Transportation Systems. Yet existing benchmarks predominantly feature relat

CAP-CoT: Cycle Adversarial Prompt for Improving Chain of Thoughts in LLM Reasoning

arXiv:2604.23270v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting has emerged as a simple and effective way to elicit step-by-step solutions from large language models (LLMs). However,

CombiMOTS: Combinatorial Multi-Objective Tree Search for Dual-Target Molecule Generation

arXiv:2604.23307v1 Announce Type: cross Abstract: Dual-target molecule generation, which focuses on discovering compounds capable of interacting with two target proteins, has garnered significant atte

Discovering Failure Modes in Vision-Language Models using RL

arXiv:2604.04733v2 Announce Type: replace-cross Abstract: Vision-language Models (VLMs), despite achieving strong performance on multimodal benchmarks, often misinterpret straightforward visual concep

Model Releases

Exploring the Secondary Risks of Large Language Models

model-releasesarxiv-cs-ai

arXiv:2506.12382v5 Announce Type: replace-cross Abstract: Ensuring the safety and alignment of Large Language Models is a significant challenge with their growing integration into critical application

Fix Initial Codes and Iteratively Refine Textual Directions Toward Safe Multi-Turn Code Correction

arXiv:2604.23989v1 Announce Type: cross Abstract: Recent work on large language models (LLMs) has emphasized the importance of scaling inference compute. From this perspective, the state-of-the-art me

Fragmented AI policy threatens US leadership as government scrambles to keep pace

AI policy fragmentation is emerging as a critical risk for Washington, and without a federal standard, a patchwork of conflicting state-level rules threatens to undermine American competitiveness. Tha

safetysiliconangle

From Optimization to Prediction: Transformer-Based Path-Flow Estimation to the Traffic Assignment Problem

arXiv:2510.19889v2 Announce Type: replace-cross Abstract: The traffic assignment problem is essential for traffic flow analysis, traditionally solved using mathematical programs under the Equilibrium

Generative Control as Optimization: Time Unconditional Flow Matching for Adaptive and Robust Robotic Control

arXiv:2603.17834v2 Announce Type: replace-cross Abstract: Diffusion models and flow matching have become a cornerstone of robotic imitation learning, yet they suffer from a structural inefficiency whe

Guided Speculative Inference for Efficient Test-Time Alignment of LLMs

arXiv:2506.04118v3 Announce Type: replace Abstract: We propose Guided Speculative Inference (GSI), a novel algorithm for efficient reward-guided decoding in large language models. GSI combines soft be

Hierarchical RL-MPC Control for Dynamic Wake Steering in Wind Farms

arXiv:2604.22797v1 Announce Type: cross Abstract: Wind farm wake steering optimization is challenging due to complex flow physics and changing conditions. This paper presents a hierarchical framework

Knowledge Lever Risk Management for Software Engineering: A Stochastic Framework for Mitigating Knowledge Loss

arXiv:2604.23257v1 Announce Type: cross Abstract: Software engineering (SE) organizations operate in a knowledge-intensive domain where critical assets -- architectural expertise, design rationale, an

Learning from Demonstration with Failure Awareness for Safe Robot Navigation

arXiv:2604.23360v1 Announce Type: new Abstract: Learning from demonstration is widely used for robot navigation, yet it suffers from a fundamental limitation: demonstrations consist predominantly of s

safetyarxiv-cs-ro

OpenPodcar2: a robust, ROS2 vehicle for self-driving research

arXiv:2604.24242v1 Announce Type: new Abstract: OpenPodcar2 is a robust, ROS2-interfaced, low-cost, open source hardware and software, autonomous vehicle platform based on an off-the-shelf, hard-canop

safetyarxiv-cs-ro

Quantifying Divergence in Inter-LLM Communication Through API Retrieval and Ranking

arXiv:2604.22760v1 Announce Type: cross Abstract: Large language models (LLMs) increasingly operate as autonomous agents that reason over external APIs to perform complex tasks. However, their reliabi

Model Releases

Reinforcement Learning with Backtracking Feedback

model-releasesarxiv-cs-ai

arXiv:2602.08377v2 Announce Type: replace-cross Abstract: Addressing the critical need for robust safety in Large Language Models (LLMs), particularly against adversarial attacks and in-distribution e

Towards Holistic Evaluation of Large Audio-Language Models: A Comprehensive Survey

arXiv:2505.15957v4 Announce Type: replace-cross Abstract: With advancements in large audio-language models (LALMs), which enhance large language models (LLMs) with auditory capabilities, these models

Transferable Physical-World Adversarial Patches Against Object Detection in Autonomous Driving

arXiv:2604.23105v1 Announce Type: new Abstract: Deep learning drives major advances in autonomous driving (AD), where object detectors are central to perception. However, adversarial attacks pose sign

UGAF-ITS: A Standards Harmonization Framework and Validation Tool for Multi-Framework AI Governance in Distributed Intelligent Transportation Systems

arXiv:2604.22789v1 Announce Type: cross Abstract: Organizations deploying AI-enabled Intelligent Transportation Systems face fragmented governance: ISO/IEC 42001 demands a certifiable management syste

UniAda: Universal Adaptive Multi-objective Adversarial Attack for End-to-End Autonomous Driving Systems

arXiv:2604.23362v1 Announce Type: cross Abstract: Adversarial attacks play a pivotal role in testing and improving the reliability of deep learning (DL) systems. Existing literature has demonstrated t

Verifying Quantized GNNs With Readout Is Decidable But Highly Intractable

arXiv:2510.08045v2 Announce Type: replace-cross Abstract: We introduce a logical language for reasoning about quantized aggregate-combine graph neural networks with global readout (ACR-GNNs). We provi

When Policies Cannot Be Retrained: A Unified Closed-Form View of Post-Training Steering in Offline Reinforcement Learning

arXiv:2604.22873v1 Announce Type: cross Abstract: Offline reinforcement learning (RL) can learn effective policies from fixed datasets, but deployment objectives may change after training, and in many

A Co-Evolutionary Theory of Human-AI Coexistence: Mutualism, Governance, and Dynamics in Complex Societies

arXiv:2604.22227v1 Announce Type: cross Abstract: Classical robot ethics is often framed around obedience, most famously through Asimov's laws. This framing is too narrow for contemporary AI systems,

How Large Language Models Balance Internal Knowledge with User and Document Assertions

arXiv:2604.22193v1 Announce Type: new Abstract: Large language models (LLMs) often need to balance their internal parametric knowledge with external information, such as user beliefs and content from

safetyarxiv-cs-cl

Is it me or X starting to look like a vibe coded mess? Polls are broken. Accounts are getting hacked. My DMs are full of phishing scams. Bas…

Gary Marcus discusses technical and security issues affecting the X platform, including malfunctioning polls, compromised accounts, and increased phishing scams in direct messages. The post appears to

safetygary-marcus--x

Learning-augmented robotic automation for real-world manufacturing

arXiv:2604.22235v1 Announce Type: cross Abstract: Industrial robots are widely used in manufacturing, yet most manipulation still depends on fixed waypoint scripts that are brittle to environmental ch

On the Properties of Feature Attribution for Supervised Contrastive Learning

arXiv:2604.22540v1 Announce Type: cross Abstract: Most Neural Networks (NNs) for classification are trained using Cross-Entropy as a loss function. This approach requires the model to have an explicit

Reliable Self-Harm Risk Screening via Adaptive Multi-Agent LLM Systems

arXiv:2604.22154v1 Announce Type: cross Abstract: Emerging AI systems in behavioral health and psychiatry use multi-step or multi-agent LLM pipelines for tasks like assessing self-harm risk and screen

Rethinking XAI Evaluation: A Human-Centered Audit of Shapley Benchmarks in High-Stakes Settings

arXiv:2604.22662v1 Announce Type: cross Abstract: Shapley values are a cornerstone of explainable AI, yet their proliferation into competing formulations has created a fragmented landscape with little

Scam Altman didn’t tell the OpenAI board that he OWNED the OpenAI Startup Fund. Altman lied in congressional testimony that he didn’t have f…

Scam Altman didn’t tell the OpenAI board that he OWNED the OpenAI Startup Fund. Altman lied in congressional testimony that he didn’t have financial gain from OpenAI. Ex-board member of OpenAI calls S

safetyelon-musk--x

The only AGI that Sam Altman is after is Adjusted Gross Income.

Gary Marcus made a critical commentary on Sam Altman's priorities, using a pun on 'AGI' (Artificial General Intelligence) to suggest that Altman's actual focus is on 'Adjusted Gross Income' rather tha

safetygary-marcus--x

The people building the most powerful technology in history cannot tell you what is happening inside their own systems. This is insane. We s…

The people building the most powerful technology in history cannot tell you what is happening inside their own systems. This is insane. We started the Torchbearer Community for exactly this reason. Re

safetyconnor-leahy--x

Coders and software engineers ONLY:

Gary Marcus, a prominent AI researcher and critic, posted a message on X (formerly Twitter) addressing software engineers and coders, likely discussing technical aspects of AI development, programming

safetygary-marcus--x

26 Apr 2026

Existential risk mongers are a small, very vocal cult with a lot of very clever online astroturfing skills. Politicians never waste a good f…

Existential risk mongers are a small, very vocal cult with a lot of very clever online astroturfing skills. Politicians never waste a good fake crisis, which is why they're perfect for Bernie to try t

safetyyann-lecun--x

26 Apr 2026

What if I told you there was a technology where 1.5 million people would die every year and injure 50 million would you sign up for that tec…