Search: “safety”

arXiv:2604.06205v1 Announce Type: cross Abstract: The growth of online platforms and user content requires strong content moderation systems that can handle complex inputs from various media types. Wh

arXiv:2610.10285v1 Announce Type: cross Abstract: In the wake of the increasingly public discussion about AI alignment, recent work has tried to propose specific AI architectures that behave safely. H

arXiv:2610.08765v1 Announce Type: new Abstract: Control barrier functions (CBFs) certify commands through an assumed dynamics model, so an abrupt, unmeasured regime change can undermine the certificat

arXiv:2610.07519v1 Announce Type: new Abstract: Automatic sign language translation (SLT) has entered consumer products, turning American Sign Language into English text for dictation, messaging, and

arXiv:2610.06171v1 Announce Type: new Abstract: Evaluating end-to-end autonomous driving under rare, safety-critical vehicle-pedestrian interactions requires photorealistic, sensor-level scenarios. Ho

arXiv:2602.03255v2 Announce Type: replace Abstract: Computer-use agents (CUAs) execute multi-stage tasks through tools, where an early unsafe decision can propagate to consequential actions. Evaluatin

arXiv:2610.01535v1 Announce Type: cross Abstract: Safety routers send each request to one of several models and are judged against the best single model. A major routing benchmark picks that comparato

arXiv:2610.01083v1 Announce Type: new Abstract: Vision-language-action (VLA) policies have demonstrated impressive capabilities in generalizable robotic manipulation, but their deployment in the real

arXiv:2609.39279v1 Announce Type: cross Abstract: Machine unlearning has emerged as a crucial mechanism for removing hazardous knowledge and enforcing safety alignment in Large Language Models (LLMs).

arXiv:2609.40177v1 Announce Type: new Abstract: Safe social navigation requires a robot to anticipate not only the future consequences of its actions, but also whether a nominal action can actually be

arXiv:2609.35415v1 Announce Type: cross Abstract: Frozen adaptive cruise control (ACC) policies can violate constraints when deployment dynamics differ from their training conditions. We propose resid

arXiv:2609.32801v1 Announce Type: cross Abstract: Embodied task planners may produce multi-step plans whose subtask dependencies and interactions with the environment create physical risks during exec

arXiv:2609.33005v1 Announce Type: new Abstract: Automated malaria diagnosis from thin blood-smear microscopy could meaningfully reduce the burden on under-resourced laboratories, but a model that maxi

arXiv:2609.31358v1 Announce Type: cross Abstract: The Model Context Protocol (MCP) provides a common interface through which AI applications discover and use external resources and tools. It allows la

arXiv:2609.30802v1 Announce Type: new Abstract: Prior research has demonstrated that the choice of prompt template during Supervised Fine-Tuning (SFT) significantly impacts the robustness of safety al

arXiv:2609.29341v1 Announce Type: new Abstract: Consumer-facing skincare AI must coordinate visual evidence, product information, tool use, and user-facing actions within explicit evidence and safety

arXiv:2603.17373v2 Announce Type: replace Abstract: Large language models are rapidly being deployed as AI tutors, yet current evaluation paradigms assess problem-solving accuracy and generic safety i

arXiv:2609.22108v1 Announce Type: cross Abstract: Learning-enabled perception is important in many autonomous systems. Unlike traditional sensors, the boundary where ML perception does or does not wor

arXiv:2609.22786v1 Announce Type: new Abstract: There are renewed efforts to build a lunar base and house a team of astronauts on the Moon for extended periods. However, important challenges remain, s

arXiv:2609.22762v1 Announce Type: new Abstract: Generative world-action models (WAMs) jointly generate future video and vehicle actions, while their action branches remain primarily optimized by exper

arXiv:2609.19366v1 Announce Type: new Abstract: Prior work has shown that internal harmfulness representations in large language models vary across risk categories, while sharing a common general harm

arXiv:2609.17865v1 Announce Type: new Abstract: Frontier models are often evaluated on how they respond to safety information once it is already in context. We study an earlier decision point: whether

arXiv:2603.23983v2 Announce Type: replace-cross Abstract: Recent advances in real-time interactive text-driven motion generation have enabled humanoids to perform diverse behaviors. However, kinematic

arXiv:2609.13657v1 Announce Type: new Abstract: Traditional recommender systems are typically trained to predict what item users will interact with next, but not why. However, offering personalized ev

arXiv:2609.11433v1 Announce Type: new Abstract: Skill adaptation frameworks based on reinforcement learning often require restrictive assumptions to maintain stability, such as fixed observations or t

arXiv:2609.06573v1 Announce Type: new Abstract: Recent studies report that automated red-teaming finds more vulnerabilities, at lower cost, than human red-teaming on standard AI safety benchmarks, and

arXiv:2609.06991v1 Announce Type: new Abstract: Recent text-to-audio-video (T2AV) models jointly generate video, speech, sound effects, and ambience from a single text prompt. This capability poses ne

arXiv:2609.07162v1 Announce Type: new Abstract: Several properties safety monitors are asked to certify, among them cross-tenant noninterference, sandbagging and evaluation awareness, are 2-safety hyp

arXiv:2609.05270v1 Announce Type: new Abstract: Artificial intelligence (AI) is transforming not only what information systems researchers design, but also how design research is conducted. Yet existi

arXiv:2609.05117v1 Announce Type: cross Abstract: Current LLM safety benchmarks largely rely on binary metrics, overlooking how models respond to harmful prompts with varying threat implicitness. We i

arXiv:2609.04022v1 Announce Type: cross Abstract: Aligning large language models (LLMs) is essential for their safe deployment. Current alignment methods mainly optimize observable responses, yet mode

arXiv:2405.18540v3 Announce Type: replace Abstract: Red-teaming, or identifying prompts that elicit harmful responses, is a critical step in ensuring the safe and responsible deployment of large langu

arXiv:2503.00610v2 Announce Type: replace-cross Abstract: Understanding how people perceive urban environments is essential for inclusive planning, yet conventional surveys are costly and difficult to

arXiv:2608.24191v1 Announce Type: cross Abstract: Urdu, the world's tenth most spoken language with 246 million speakers, remains almost entirely absent from mainstream LLM safety evaluation and nine

arXiv:2606.18632v2 Announce Type: replace Abstract: Embodied Foundation Models (EFMs) integrate multimodal understanding, future-state reasoning, and executable robot actions. Yet their safety alignme

arXiv:2608.17202v1 Announce Type: new Abstract: Safety alignment in open-weight language models is trivially removable: abliteration projects a refusal-mediating direction out of the weights in minute

arXiv:2608.13345v1 Announce Type: new Abstract: Artificial Intelligence (AI) safety systems combine character shaping (e.g., Reinforcement Learning from Human Feedback [RLHF], Constitutional AI), whic

arXiv:2608.11498v1 Announce Type: new Abstract: Natural-language-based scenario generation offers an intuitive means of describing rare and complex driving interactions, yet it is still uncertain whet

arXiv:2509.13450v3 Announce Type: replace Abstract: We introduce SteeringSafety, a benchmark for evaluating representation steering methods across nine safety perspectives spanning 18 datasets. While

arXiv:2608.09542v1 Announce Type: cross Abstract: Large reasoning models (LRMs) achieve remarkable success on complex tasks but remain vulnerable to harmful prompts that induce unsafe outputs. Recent

arXiv:2608.08077v1 Announce Type: new Abstract: Theory of Space framework (ToS) assesses the spatial understanding of curiosity-driven Vision-Language Models (VLMs) under partial observability. As AI

arXiv:2608.08542v1 Announce Type: new Abstract: Model merging has become the default way to give an aligned language model new skills without retraining: a practitioner folds task vectors from math, c

arXiv:2608.05560v1 Announce Type: cross Abstract: Timely anticipation of physical hazards is essential for real-world safety, yet existing MLLM evaluations focus on harmful content or general risks, l

arXiv:2508.16440v2 Announce Type: replace-cross Abstract: Urban Air Mobility (UAM) envisions the widespread use of small aerial vehicles to transform transportation in dense urban environments. Howeve

arXiv:2608.05086v1 Announce Type: new Abstract: Language models differ in how safely they behave and these differences are measured by safety benchmarks. But aggregated benchmark scores are hard to tr

arXiv:2608.02941v1 Announce Type: new Abstract: We audit five frontier large language models on native Bangla derogatory speech (gali) across six protocols to test a single hypothesis: Comprehension-C