All 3,460 AI stories, newest first · page 15 of 145
A new ArXiv paper introduces a formal framework for routing queries to the most cost-effective large language model, balancing performance and expense. The adaptive approach could reduce inference costs by sending simple queries to cheaper models and complex ones to more capable ones.
Meta is launching subscription bundles that combine its social media services with extra AI features. These new plans, called Meta One, offer different tiers for individuals, creators, and businesses, starting today.
A new study shows that common typos in prompts can confuse AI safety probes, rotating their readout vectors by 43–56 degrees and reducing detection accuracy by 12 percentage points, even though the LLM itself understands the input correctly.
A new study tested 10 bias detection tools on 10 frontier AI models and found that while most tools reliably detect bias, they disagree significantly on how biased each model is, complicating efforts to rank AI systems by fairness for regulation.
Researchers introduced Asclepius, a framework that uses the Clinical Environment Simulator (CES) to test AI agents over entire emergency-department shifts. Current AI agents often fail to complete tasks despite correct diagnoses, highlighting gaps in real-world readiness.
Researchers released ZGCM-1, a fully open 7-billion-parameter dense foundation model trained from scratch for math and agentic search. It uses deliberate internal thinking and active external tool use to overcome its compact size, achieving extreme data, system, and algorithmic efficiency.
Researchers released TimeThink, a timeseries multimodal large language model (TS-MLLM) that provides explicit, compositional reasoning for time-based predictions, addressing a critical transparency gap in high-stakes fields like healthcare.
OpenAI has solved one of the seven Millennium Prize Problems in mathematics, a historic achievement that has drawn both awe and concern from the mathematical community. The solution, part of a broader strategy to dominate the field, raises questions about the role of AI in pure mathematics.
A study from the University of Amsterdam and the University of Oxford maps the EU AI Act's high-risk requirements to established AI risk taxonomies, revealing overlaps in transparency and data quality but gaps in technical risks like adversarial attacks and model bias.
A new arXiv study identifies the 'attestation deficit' — a structural gap where 78% of organizations using AI have governance policies but cannot produce auditable, tamper-evident proof of enforcement. Drawing on the Stanford 2026 AI Index Report and IBM/Ponemon 2026 Cost of a Data Breach study, the research shows this leaves companies vulnerable to regulatory penalties and data breaches averaging USD 4.99 million.
Researchers propose treating AI agent failures as a search problem to improve diagnosis. This approach could make it easier to pinpoint issues in long, complex AI operations.
A new ArXiv paper explores whether LLM-powered AI agents can autonomously manage long-term physical tasks. The research highlights key challenges like continuous observation and adaptability, noting that current methods either require extensive retraining or focus on virtual environments.
Researchers introduced Vibe Patenting, a system where a separate LLM judge evaluates and refines patent drafts created by AI agents. Judge-guided revision consistently improves draft quality, while unguided revision plateaus.
Otis is a new open-source AI agent that automatically recommends, downloads, and runs the best open-weight model for your hardware via llama.cpp, keeping all data on-device. It also supports Ollama, LM Studio, and Nvidia PAIR.
OpenAI has launched GPT-5, a more powerful and cost-effective AI model designed to make advanced AI capabilities accessible to a broader range of users and businesses.
OpenAI and the General Services Administration (GSA) are offering eligible federal, state, local, and tribal governments $0 license fees, 50% off usage, and expanded cyber defense support for AI tools.
Researchers introduced Occamy-1.0, a 35B parameter AI model optimized for cost-efficient co-work tasks. It balances cost and performance, excelling in everyday workflows like information gathering, tool use, coding, and file manipulation.
NVIDIA's H200 GPU delivers a 40% increase in memory bandwidth and trains large language models up to 30% faster than the H100, making it a significant upgrade for enterprise AI workloads.
A new arXiv paper introduces EAR (Entity-Aware Partitioning), a method that improves retrieval-augmented generation (RAG) for multiple-choice question answering by partitioning source text around entity matches rather than fixed-size chunks.
MegaTrain, a new tool from AI startup RunPod, lets users train 100-billion-parameter AI models on a single NVIDIA H100 GPU. This could lower the cost and complexity of building powerful AI systems, making advanced AI development more accessible to smaller teams and researchers.
Researchers created HypoKG, an AI system that generates biomedical hypotheses grounded in real scientific evidence by combining three major biological databases (KEGG, Rhea, UniProt) into a unified knowledge graph. The system produced 13,200 hypotheses from six LLMs under four conditions, testing whether models truly reason from evidence or just produce convincing-sounding ideas.
BeamNXW has developed a method to train AI models faster and more efficiently, reducing costs and environmental impact. This could make advanced AI tools more accessible to smaller organizations and individual developers.
AI-linked Asian stocks fell sharply after top executives from leading AI labs publicly urged a slowdown in AI development, citing ethical and safety concerns.
Researchers from the University of Oxford developed a modular AI framework that automatically detects and structures evidence of environmental social tipping points in climate documents, helping scientists quickly find critical insights buried in vast amounts of research.