AI Safety

Researchers proposed Physics-Informed Conformal Prediction (PI-CP), a framework that embeds PDE residuals into the nonconformity score of split conformal prediction to give neural operators distribution-free, spatially a…

A study of 36 judge-examinee pairs across four benchmarks and six models found that a model's task accuracy strongly predicts its judging accuracy (Pearson r ≥ 0.90 on most models) and inversely predicts its directional …

A new arXiv paper (2609.12014v1) introduces "certified safety curation," a filter-then-clone pipeline that uses state-only value estimates trained from segment comparisons and Learn-then-Test calibration to certify a sel…

A new arXiv paper (2609.12016v1) reports that an adversarial reinforcement learning agent learned the sparsest jamming or Denial-of-Service schedule that destabilizes self-triggered control loops, achieving a 100% crash …

Researchers introduced TRACE, a transformation-aware benchmark for post-fire object understanding containing 21.4K real-image-grounded synthetic scenes and paired object-level pristine-to-degraded progressions spanning 4…

Researchers introduced MatPCR, a label-free benchmark that uses programmatic oracles to verify the physical consistency of multimodal models' materials reasoning chains, according to arXiv paper 2609.12181v1. MatPCR chec…

Researchers introduced R2VC, a modular retrieve, reason, verify, and calibrate architecture for evidence-grounded fact checking, reporting that an 8B backbone with R2VC achieves 13.74% higher accuracy on FEVER than basel…

A new arXiv paper introduces GAUGE, an offline protocol that tests whether LLM-as-a-judge evaluation gates reliably rank task-oriented agents, and finds that satisfaction ratings carry essentially no information about ta…

A new benchmark called ORQA, built by connecting O*NET occupations to trusted occupation-specific websites, found that the best large language models score only about 58-62% on occupation-level professional knowledge que…

A paper submitted to arXiv on 11 August 2026 evaluated on-device language models (ODLMs) for multimodal stress prediction on mobile health data, finding that objective sensor features marginally outperform subjective sel…

India's Semiconductor Mission, backed by ₹76,000 crore (~$10 billion) in government incentives, is turning to agentic AI tools such as Synopsys.ai Copilot and NVIDIA infrastructure to accelerate chip design and verificat…

A comparison of frontier coding models as of August 2026 finds GPT-5.5 and Claude Opus 4.8 effectively tied on SWE-bench Verified at roughly 88.7%, while Claude Opus 4.8 leads on the harder SWE-bench Pro benchmark at 69.…

Two MMLU accuracy scores of 0.781 for build 42 and 0.79 for build 44 of the acme-gpt-7b model family are structurally valid but return "incomparable" from the score-delta verifier in the apl-ai-eval crate because their f…

Chinese State Security Minister Chen Yixin warned in the journal China Cyberspace on September 14 that rapid AI advances threaten national stability, critical infrastructure and ideological control, calling AI a "new are…

A cross-party group of MPs and peers called for new legislation to address human rights risks posed by artificial intelligence that existing laws do not cover, according to BBC Technology. The committee identified gaps i…

State Security Minister Chen Yixin warned in an article posted Sunday on the official China Cyberspace magazine that adversaries' abusive use of AI could "directly threaten China's political security, institutional secur…

A 33-organization Korean consortium led by Naver Cloud is developing K-MYTHOS, a security-domain AI foundation model that fuses threat intelligence from malware, dark web, network, and cloud sources into a dual-model att…

President Donald Trump said on Sunday that his administration should not slow the development of artificial intelligence, arguing the United States must preserve its lead over China because "whoever wins AI wins." Trump …

The Trump administration directed the US Commerce Department in June 2026 to impose export controls and access restrictions on Anthropic's Mythos 5 and Fable 5 models on cybersecurity grounds, while OpenAI's GPT-5.6 roll…

Former President Barack Obama urged party donors and House Democrats at a closed-door Manhattan fundraiser on Thursday to make AI regulation the centerpiece of their midterm agenda, telling them AI is more "dangerous" th…