Knowledge catalogue
When you hear AI 'safety' you should hear 'censorship' and 'control' instead. All of us surveilled and spied by safeguards of loving grace. Today it's intelligent Terms of Service control. You can't d
arXiv:2606.08969v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly used for medical summarization, but their outputs can omit medically important information and introduce
fable’s safety guardrails broken within an hour. raise your hand if you are surprised. We tested Anthropic’s new @claudeai Fable 5. It did not fail like an ordinary jailbreak. It failed more quietly.
arXiv:2606.09038v1 Announce Type: new Abstract: Large Language Models (LLMs) have enabled increasingly personalized interactions by adapting to users' preferences, contexts, and long-term histories. H
arXiv:2606.08860v1 Announce Type: new Abstract: Temporary work-zone speed limits are communicated through visually inconsistent signage and are often missing from digital maps, creating safety risks f
arXiv:2511.20158v2 Announce Type: replace Abstract: While continual visual instruction tuning (CVIT) has shown promise in adapting multimodal large language models (MLLMs), existing studies predominan
Not just foreseeable, but foreseen and called out. We ran a campaign to get them out of the first ever AI Safety Summit. We won that one. Yet most of the field kept licking the boot of the companies a
Safety officials gained detailed data on the impact of large rocket explosions when Blue Origin's New Glenn rocket experienced an anomaly during a hot-fire test at Cape Canaveral on May 28, causing a
David Jeans / Reuters: Sources and docs detail defense tech startup Shield AI's struggles to overcome years of technical hitches and safety concerns with its V-BAT autonomous drone — A year ago, Ryan
arXiv:2602.06911v2 Announce Type: replace-cross Abstract: As increasingly capable open-weight large language models (LLMs) are deployed, improving their tamper resistance against unsafe modifications,
arXiv:2606.04035v1 Announce Type: cross Abstract: We present a systematic study of domain-dependent safety behavior in open-weight LLMs: 7 standardized experiments across 7 ethical domains, testing 5
OpenAI outlines initiatives and commitments to enhance youth safety while creating educational and economic opportunities for young people globally. The effort emphasizes OpenAI's role in responsible
arXiv:2606.00773v1 Announce Type: new Abstract: Vision-language-action (VLA) benchmarks measure whether a policy completes a requested manipulation task, but binary success can hide safety-relevant tr
AI safety can't happen behind closed doors! Super cool to see that the @AISecurityInst is releasing its evals, datasets, and models in the open on @huggingface, so researchers everywhere can scrutiniz
arXiv:2605.29340v1 Announce Type: new Abstract: In this paper, we discuss question-answer dataset for LLM safety evaluation, with a focus on illegal activities. Specifically, on the basis of manual an
arXiv:2605.29801v1 Announce Type: new Abstract: Modern open-world agents such as OpenClaw exhibit powerful cross-environment execution capabilities yet introduce broad new safety risk sources. Meanwhi
Reuters: Former Tesla data labelers say FSD relies on laborious mapping for hazards; crash data analysis shows Tesla exaggerates FSD's safety via flawed methodology — Tesla says its Full Self-Driving
arXiv:2605.28552v1 Announce Type: new Abstract: As automated vehicles (AVs) increasingly share roadways with human-driven vehicles (HDVs), understanding how pedestrians respond to different vehicle ty
arXiv:2605.26954v1 Announce Type: new Abstract: Safety evaluation of Large Language Models (LLMs) has largely focused on high-resource languages, leaving low-resource languages critically underserved.
Jared Perlo / NBC News: Illinois Legislature passes SB 315, a bill requiring annual independent third-party safety audits of leading AI companies; the bill heads to the governor's desk — The measure,
arXiv:2605.27117v1 Announce Type: new Abstract: AI safety is still largely framed as alignment: training models to follow human preferences, safety policies, and normative constraints. That framing ha
arXiv:2505.11063v3 Announce Type: replace Abstract: LLM-based agents solve complex tasks through iterative reasoning, tool use, and environment interaction, where each intermediate thought directly sh
Looking forward to speaking at BloombergTech next week with @shiringhaffary to discuss AI safety and our research progress at @LawZero_! Is AI development progressing too quickly? @business' @shiringh
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