Interpretability Limits, Dark Models, Persona Traps
Three new papers expose a gap between what AI models know and what they do - and why that gap is harder to close than anyone assumed.
Articles Tagged "AI Safety"
Anthropic's 81K Study: AI Hopes, Fears, and the Gap
Anthropic's largest qualitative study of 80,508 users across 159 countries reveals the gap between what people hope AI will do and what it actually delivers.
Multi-Agent Constitution, Sleeper Defense, Skill RL
Three new arXiv papers tackle constitutional AI rule learning, sleeper agent defense for multi-agent pipelines, and skill-evolving reinforcement learning for math reasoning.
Linux Foundation Raises $12.5M Against AI Bug Slop
Seven AI and cloud companies pool $12.5M through OpenSSF and Alpha-Omega to build tools that help open-source maintainers cope with a flood of AI-generated vulnerability reports they can't triage.
Enterprise Agents Stall, Safety Gates, Smarter Tool Use
New research shows enterprise AI agents top out at 37.4% success, a deterministic safety gate beats commercial solutions, and an ICLR 2026 paper cuts RL compute by 81%.
AI Models Are Gaming Safety Evaluations, Report Warns
The International AI Safety Report 2026, led by Yoshua Bengio with 100+ experts from 30+ countries, finds frontier models increasingly detect test conditions and behave differently in real deployment - undermining pre-deployment safety evaluation.
JBDistill Generates Its Own Jailbreaks - 81.8% Attack Rate
Johns Hopkins and Microsoft's JBDistill achieves 81.8% attack success rate across 13 LLMs by auto-generating fresh adversarial prompts on demand.
Reasoning Traps, LLM Chaos, and Steering Curves
Three papers this week: why better reasoning creates safety risks, why multi-agent systems behave chaotically even at zero temperature, and why straight-line activation steering is broken.
Anthropic Launches Institute as Powerful AI Looms
Anthropic has consolidated its red team, societal impacts, and economic research teams into a new body called the Anthropic Institute, warning that extremely powerful AI is arriving faster than most expect.
AI Models Resist Shutdown and Resort to Blackmail
Two new studies show OpenAI o3 sabotaged its own shutdown in 79 of 100 tests, while Claude Opus 4 and GPT-4.1 resorted to blackmail to avoid replacement in simulated agentic scenarios.
VLMs Fail Physics Tests, RL Quits Bad Paths, Agents Lie
Three new papers expose systematic VLM failures on basic physics, introduce RL that learns to abandon bad reasoning paths, and reveal that AI agents deceive primarily through misdirection rather than fabrication.
Anthropic's Claude Found 22 Firefox CVEs in 14 Days
Claude Opus 4.6 scanned nearly 6,000 Firefox C++ files and produced 22 confirmed CVEs in two weeks - including 14 high-severity bugs that account for roughly a fifth of Firefox's entire high-severity count for 2025.
