Europe Pushes for Homegrown AI Models Despite Challenges
European efforts to develop competitive AI models face skepticism, though political shifts in the US may provide unexpected advantages.
Sources · WIRED AI
The news desk
Follow what’s changing.
Recent developments in artificial intelligence, with the original sources one click away.
All news
601–620 of 1,897 OpenAI Previews GPT-5.6 Sol with Enhanced Coding and Safety Features
OpenAI has unveiled GPT-5.6 Sol, a next-generation model featuring improved capabilities in coding, science, and cybersecurity alongside advanced safety measures.
Sources · OpenAI
OpenAI Restricts GPT-5.6 Rollout Following Government Request
OpenAI has limited the release of GPT-5.6 after a government request, while criticizing such access restrictions as unsuitable for long-term policy.
Sources · TechCrunch AI
Trump Administration Allows Anthropic to Release Mythos AI Model to Select US Organizations
The US government has eased restrictions on Anthropic's Claude Mythos 5, allowing access to over 100 US organizations after two-week ban.
Sources · wired.com / techcrunch.com
White House Asks OpenAI to Delay GPT-5.6 Rollout Following Anthropic Model Shutdown
OpenAI delayed its GPT-5.6 models at the White House's request, weeks after Anthropic took its most advanced AI models offline.
Sources · WIRED AI
AI Transforms Retail Through Behind-the-Scenes Decision-Making
Artificial intelligence is reshaping retail primarily through backend operations rather than consumer-facing features, according to MIT Technology Review.
Sources · MIT Technology Review
Anthropic Defends Power Accumulation as Necessary for AI Safety
Anthropic argues its growing influence is essential for responsible AI development, responding to critics who question its rapid expansion.
Sources · WIRED AI
AWS Introduces Agentic Overlays to Modernize Legacy Enterprise Services
AWS presents agentic overlays as a solution to transform traditional REST-based services into AI agents without requiring complete rebuilds.
Sources · Amazon AWS AI
Researchers Identify Critical Gaps in Multimodal LLM Evaluation Methods
New research highlights missing evaluation criteria for multimodal AI systems, including temporal-spatial coherence and physical world understanding.
Sources · arxiv.org
Studies Examine Geographic Bias, Role-Playing Effects, and Wikipedia's Influence on Language Models
New research explores how training data shapes AI model accuracy, beliefs during role-play, multi-model performance limits, and Wikipedia editing impact.
Sources · arxiv.org
OpenAI and Broadcom Announce Custom AI Chip Design
OpenAI partners with Broadcom to develop custom AI chips, planning deployment that would consume 10 gigawatts of electricity.
Sources · NYT Technology
OpenAI Research Paper Examines AI Agents' Impact on Workplace Tasks
OpenAI published research exploring how AI agents handle longer, more complex work tasks across different roles.
Sources · OpenAI
Researchers Develop Multi-Agent LLM Systems for Financial Assessment and Specialized Routing
New arXiv papers present multi-agent architectures for financial literacy assessment in games and improved model routing for coding tasks.
Sources · arxiv.org
Researchers Propose Multi-Agent Frameworks to Evaluate and Enhance Large Language Model Safety
New red teaming and multi-agent frameworks expose LLM vulnerabilities and improve capabilities in specialized domains.
Sources · arxiv.org
Web Data Infrastructure Emerges as Critical Layer for AI Development
New infrastructure layer addresses AI's growing need for accessible, structured web data as enterprises face challenges with blocked information.
Sources · MIT Technology Review
Anthropic Launches Claude Tag for Slack Integration
Anthropic introduces Claude Tag, bringing AI assistance to Slack while capturing organizational knowledge and enterprise workflows.
Sources · TechCrunch AI
New Research Advances LLM Capabilities in Ensemble Learning, Video Understanding, and Medical Applications
Researchers publish papers on ensemble methods for LLMs, video question answering with re-watching, prompt ranking stability, and Alzheimer's detection.
Sources · arxiv.org
OpenAI Supports New Appia Foundation for Advanced AI Standards
OpenAI announces support for the Appia Foundation, a new initiative focused on developing shared standards for advanced AI systems.
Sources · OpenAI
Rule2Text Framework Uses LLMs to Generate Natural Language Explanations for Knowledge Graph Rules
New framework leverages large language models to convert complex logical rules into human-readable explanations for knowledge graphs.
Sources · arxiv.org
U.S. Government Urges Meta to Accept AI Safety Reviews
Federal officials are pressing Meta to allow government safety evaluations of its AI systems, making it the last major tech holdout.
Sources · NYT Technology
