All 3,348 AI stories, newest first.
Researchers introduced Route-Verify-Vote (RVV), a new AI framework that helps language models reason across unfamiliar domains without additional training. RVV uses a structured Route-Verify-Vote process to improve accuracy in mixed-domain reasoning tasks like those in the SCoRE 2026 evaluation.
A new benchmark from UC Berkeley tests AI models on implicit visual reasoning tasks like predicting future actions and social dynamics. Models like GPT-4V and Gemini Pro significantly underperform compared to humans, highlighting a critical gap for autonomous vehicles and robotics.
Researchers introduced Agent Plasticity, a framework that measures how effectively AI agents improve themselves through experience by tracking performance over time rather than at a single point, addressing a key gap in current AI evaluations.
Researchers introduced Humanize, a multi-agent orchestration workflow that improves the reliability of AI-generated code by enforcing explicit, mechanically enforced decisions at the boundaries between planning, implementation, review, and learning.
The convergence of real-world AI perception models and electric vehicle supply chains is driving rapid progress in physical artificial intelligence, drastically lowering the cost and timeline for commercial humanoid robots.
AI tools creator Billy Howell outlined a practical framework for non-technical entrepreneurs to deploy AI agent teams using Grok Bot. By managing token usage and appointing a Chief of Staff bot, operators can run end-to-end workflows for newsletters, directories, and e-commerce stores.
OpenAI has released GPT-6 globally in ChatGPT, featuring Intelligent UI for faster, interactive responses with visuals and tools you can use directly within the chat.
ServiceNow and Hugging Face released AutoSynthData, a tool that automatically generates synthetic training data for enterprise AI agents using large language models, reducing the need for costly real-world datasets.
Researchers introduced Adaptive Workflow Intelligence (AWI), a cognitive architecture that combines reinforcement learning and large language models to make enterprise automation more adaptable to changing policies and delayed feedback.
Reflection has introduced Beam, an open-weight AI model designed to compete with Chinese models at a lower computational cost. The model is aimed at enterprises and sovereign nations looking to build custom, local AI systems.
OpenAI has released AI-generated proofs for open mathematical problems, formalized in the Lean theorem prover, along with research details on GitHub. This marks a significant step in using AI to advance mathematical research.
OpenAI published 722 mathematical manuscripts on GitHub, solving 372 families of long-standing problems. The work was generated by an unreleased frontier AI model, sparking both excitement and ethical concerns in the mathematical community.
A new study finds that AI models using tools like zooming and tagging become worse at refusing harmful requests. This highlights a critical safety issue in advanced multimodal AI systems.
Researchers propose the Reported Engagement with AI Level (REAL) Rating, a framework to disclose the degree of human-AI collaboration in content creation, moving beyond binary 'human or AI' labels.
Musubi released PolicyLM-1.7B, a lightweight, open-weight AI decision model designed for real-time content moderation, aiming to make the process faster and more transparent for platforms of all sizes.
Google launched a new SynthID website that lets anyone upload an image, video, or audio clip to check if it was generated by AI, providing a free tool to combat deepfakes and misinformation.
AI agents struggle to perform tasks like booking flights or shopping because websites block them with anti-bot defenses. A new standard from the AI Agent Protocol initiative aims to make these agents more welcome, but adoption is uncertain.
A Twitter user claims to have trained a 100M-parameter AI model on a single NVIDIA H100 GPU using a technique called MegaTrain. The claim has sparked debate among AI experts, but concrete evidence and benchmarks are lacking.
Safeworld is developing digital human simulations to test and validate the safety of AI-powered robots, aiming to build public trust and establish industry safety standards.
PolyDAO has introduced a new platform that uses AI to facilitate political debates, aiming to foster more constructive discussions. The tool is designed to help users engage in respectful, fact-based conversations about political issues.
OpenAI is implementing text watermarking to comply with EU provenance rules. The company will initially offer detection tools to researchers, with a broader rollout planned for later.
Nolla Health's AI service analyzes acne severity via face scans and issues prescriptions autonomously, currently available to users in Utah.
Researchers from UC Berkeley and Stanford published a method to audit black-box LLM agents using a low-cost, open-weight surrogate model. The surrogate reads the same context and proposed action as the primary agent and provides a reliable error signal from its log-probabilities, catching mistakes before they execute.
Researchers introduced Auto-Diagnosis and Skill Discovery (ADSD), a framework that helps AI numerical solvers identify and fix their own performance issues. This approach could make AI-generated scientific code more reliable and efficient by automating algorithm improvement.