Knowledge catalogue
arXiv:2605.24154v1 Announce Type: new Abstract: Current safety alignment of foundation models largely follows a one-size-fits-all paradigm, applying the same refusal policy across users and contexts.
arXiv:2601.21094v2 Announce Type: replace-cross Abstract: Safe Reinforcement Learning (RL) algorithms are typically evaluated under fixed training conditions. We investigate whether training-time safe
I once asked a bunch of e/accs how much damage was an acceptable risk relative to take (any) safety precautions. none answered. we may be about to find out. One implication of the below is that we rea
New York Times: Q&A with Sundar Pichai on the future of Google Search, Google's place in the AI race, public skepticism toward AI, AI agents, AI safety, TPUs, and more — After a busy Google I/O, the c
arXiv:2605.21609v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly embedded in adolescent digital environments, mediating information seeking, advice, and emotionally sensit
arXiv:2605.22185v1 Announce Type: new Abstract: Recent advancements in Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in general visual understanding. However, thei
Yeah, I don't think I've ever met someone who has directly worked for an extended period on safety or security at an AI company who thinks things are fine readiness wise or incentive wise etc. https:/
arXiv:2605.21138v1 Announce Type: new Abstract: Smoothed implicit contact dynamics enables gradient-based planning and control for contact-rich tasks without predefined mode sequences. However, safety
Brendan Bordelon / Politico: OpenAI's Chris Lehane says he is pursuing “reverse federalism”, lobbying blue states to pass AI safety laws and create a de facto US standard, as DC dithers — OpenAI's eff
arXiv:2605.17971v1 Announce Type: cross Abstract: Despite rigorous safety alignment, Large Language Models (LLMs) remain vulnerable to jailbreak attacks. Existing black-box methods often rely on heuri
arXiv:2603.04459v3 Announce Type: replace-cross Abstract: The rapid expansion of research in LLM safety presents challenges in tracking advancements, making benchmarks important evaluation infrastruct
arXiv:2605.15509v1 Announce Type: new Abstract: While Isaac Lab provides massive parallel UAV simulation, OmniSafe and safe-control-gym provide constrained-RL benchmarks, and CBFKit provides control-b
Transformer: OpenAI's disavowal of a liability shield in Illinois SB 3444 bill and endorsement of a stronger SB 315 suggest it is open to meaningful AI safety legislation — Transformer Weekly: US-Chin
arXiv:2605.07250v1 Announce Type: cross Abstract: Recent advancements in visual context compression enable MLLMs to process ultra-long contexts efficiently by rendering text into images. However, we i
arXiv:2605.05995v2 Announce Type: replace-cross Abstract: The safety alignment of Large Language Models (LLMs) remains vulnerable to Harmful Fine-tuning (HFT). While existing defenses impose constrain
We also had three third-party AI safety organizations provide feedback on our analysis: @redwood_ai, @apolloaievals, @METR_Evals. You can find @redwood_ai's report here: https://blog.redwoodresearch.o
arXiv:2605.04085v1 Announce Type: cross Abstract: Objectives: Large language models (LLMs) are increasingly used for clinical text summarization, yet structured methods to assess associated patient sa
Many, many OpenAI employees quit over safety concerns, including @DKokotajlo, William Saunders, @sjgadler, etc as well @Janleike. The founders of Anthropic such as @DarioAmodei and @jackclarkSF may ha
Sam Tobin / Reuters: Meta challenges Ofcom in UK High Court over the Online Safety Act, which calculates levies based on global, not UK, revenue, in a case scheduled for October — Facebook and Instagr
arXiv:2508.04204v2 Announce Type: replace Abstract: Large Reasoning Models (LRMs) have demonstrated impressive performance in reasoning-intensive tasks, but they remain vulnerable to harmful content g
Jay Peters / The Verge: Musk v. Altman: Mira Murati testifies that Sam Altman lied to her about the safety standards for a new OpenAI model and that he made her work more difficult — OpenAI's former C
arXiv:2605.02544v1 Announce Type: cross Abstract: The widespread adoption of machine learning in critical applications demands techniques to mitigate high-consequence errors. Our method utilizes a dua
arXiv:2605.00689v1 Announce Type: new Abstract: As Large Language Models (LLMs) are increasingly deployed in cross-linguistic contexts, ensuring safety in diverse regulatory and cultural environments
Diana Novak Jones / Reuters: New Mexico child safety trial: New Mexico asks a judge to declare Meta a public nuisance and to order it to pay $3.7B and overhaul its apps to protect children — The U.S.
Ben Cohen / Wall Street Journal: A profile of BlackBerry's QNX division, whose operating system controls safety features in 275M cars and accounts for half of BlackBerry's revenue — John Wall has spen
arXiv:2604.26959v1 Announce Type: cross Abstract: Integrating large language models (LLMs) into patient-facing healthcare systems offers significant potential to improve access to medical information.
arXiv:2601.19768v3 Announce Type: replace Abstract: Large language models (LLMs) are increasingly paired with activation-based monitoring to detect and prevent harmful behaviors that may not be appare
Dear @elonmusk, If you still genuinely care about AI safety, you can’t let the Trump administration leave the AI industry almost entirely unregulated. You just can’t. - Gary The judge just instructed
arXiv:2604.26167v1 Announce Type: cross Abstract: Recent work has shown that a model's input word embeddings can serve as effective control variables for steering its behavior toward outputs that sati
arXiv:2602.15853v2 Announce Type: replace-cross Abstract: We propose a lightweight explainable guardrail (LEG) method to detect unsafe prompts. LEG uses a multi-task learning architecture to jointly l
arXiv:2604.23854v1 Announce Type: new Abstract: The application of Deep Learning in medical diagnosis must balance patient safety with compliance with data protection regulations. Machine Unlearning e
arXiv:2511.08484v2 Announce Type: replace Abstract: We propose patching for large language models (LLMs) like software versions, a lightweight and modular approach for addressing safety vulnerabilitie
arXiv:2604.23859v1 Announce Type: new Abstract: With spotforecast2-safe we present an integrated Compliance-by-Design approach to Python-based point forecasting of time series in safety-critical envir
Renju Jose / Reuters: Australia's eSafety Commissioner issues transparency notices to Roblox, Microsoft's Minecraft, and other online gaming platforms to detail child safety measures — Australia's int
arXiv:2601.22737v2 Announce Type: replace Abstract: The robust safety of Vision-Language Large Models (VLLMs) against joint multilingual and multimodal threats remains severely underexplored. Current
We think ControlAI can turn $50M / year into a 10% chance of banning ASI. Most of the AI safety community has been far too coy about extinction risk. We're not. It's not that complicated: AI smarter t
arXiv:2509.05367v4 Announce Type: replace-cross Abstract: Large Language Model safety alignment predominantly operates on a binary assumption that requests are either safe or unsafe. This classificati
arXiv:2510.14133v2 Announce Type: replace Abstract: Agentic AI systems, which leverage multiple autonomous agents and large language models (LLMs), are increasingly used to address complex, multi-step
arXiv:2604.12500v1 Announce Type: new Abstract: Specification gaming under Reinforcement Learning (RL) is known to cause LLMs to develop sycophantic, manipulative, or deceptive behavior, yet the condi
Anthropic believes that good transparency legislation needs to ensure public safety and accountability for the companies developing this powerful technology, not provide a get-out-of-jail-free card ag
arXiv:2604.10299v1 Announce Type: cross Abstract: Large Vision-Language Models (LVLMs) rely on attention-based retrieval of safety instructions to maintain alignment during generation. Existing attack
Given the extremely high rate of recidivism, this is important for community safety Murder registry is a good idea from Elon. People should know if they are close to a murderer so proper precautions c
arXiv:2604.02022v2 Announce Type: replace Abstract: Evaluating the safety of LLM-based agents is increasingly important because risks in realistic deployments often emerge over multi-step interactions
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
Tesla V14.3 self-driving review. The point releases will bring polish. V15 will far exceed human levels of safety, even in completely unsupervised and complex situations. 600 miles in with FSD v14.3 a
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