Latest AI (Artificial Intelligence) News
Anthropic Sues Pentagon Over AI Supply Chain Risk Label
Anthropic is in a legal battle with the Pentagon after Defense Secretary Pete Hegs labeled it a supply chain risk, raising fears of government surveillance using AI. Experts warn of privacy threats from AI paired with Trump administration data collections, despite ongoing use of Anthropic's Claude AI.
OpenAI faced backlash for its Pentagon deal but clarified no domestic surveillance use.
Berkeley Lab Unveils AQuaRef: AI Tool Revolutionizes Protein Structure Prediction
Scientists at Lawrence Berkeley National Lab developed AQuaRef, combining AI and quantum calculations for faster, more accurate protein structures at lower cost. The tool excelled in 71 tests and mapped the elusive DJ-1 protein linked to Parkinson’s, aiding drug design and crop productivity.
Collaborators include Carnegie Mellon, University of Wrocław, and others.
Ami Labs Launches with $1B Funding for AI World Models
French startup Ami Labs raised $1B at $3.5B valuation to develop world models, advancing beyond current LLMs for industrial uses like robots and autonomous vehicles. Backed by President Macron, it targets health, industry, and partners with Meta for AR applications.
Focus shifts to complex reasoning over chatbots.
DeepMind Marks 10 Years of AlphaGo, Eyes AGI Path
DeepMind reflects on AlphaGo's legacy, crediting its methods for AlphaFold's protein folding breakthrough, Nobel-winning impact on 200M proteins. New AI co-scientist debates hypotheses, aiding antimicrobial resistance and fusion energy research.
Gemini world models plus search tools key to AGI.
AI Emerges as Key Tool in Cancer Neuroscience Evolution
A Science Signaling perspective posits AI as essential for decoding complex cancer-neuron networks in oncology and neuroscience. Machine learning identifies hybrid tumor cells, predicts metastases, and enables network-based therapies.
Future uses include pain management, protein folding at synapses, and agentic AI for hypothesis generation.
USC AI Learns Beyond Training Data in Breakthrough Study
USC Viterbi research shows AI can boost performance to 96% on untrained data via iterative recompilation, defying data limitations. Accepted at IEEE SoutheastCon 2026, this method dramatically improves extrapolation.
Lead researcher surprised by gains from simple persistence.
China's Robot School Trains Humanoids for Everyday Tasks
Euronews reports on a Chinese facility where humanoid robots learn daily activities, highlighting rapid AI robotics progress. Featured in March 11 morning bulletin amid global AI themes.
Ties into broader sustainable supercomputing visions.
Gigabyte Shares AI Democratization Vision via Green Supercomputers
Taiwanese firm Gigabyte outlines plans for sustainable supercomputers to make AI accessible, focusing on energy efficiency. Part of Euronews coverage on EU policy and tech sustainability.
Aligns with global push for green AI infrastructure.
Ex-OpenAI Researcher Releases Tool Sparking AGI Research Automation Debate
Former OpenAI researcher unveils new AI system, fueling discussions on AI surpassing humans in research and triggering intelligence explosion. Hypothesis projects automated AI R&D accelerating progress.
Ties into broader AGI rollout preparations.
Nvidia Plans Open-Source AI Agent Platform Nemo Claw
Nvidia to launch Nemo Claw, enabling enterprises to deploy custom AI agents, positioning as infrastructure for AI economy. Aligns with crypto and enterprise shifts to agentic AI.
BMNR updates highlight Ethereum's role in this layer.
2026 MIT AI Conference Explores Practical AGI Horizons
MIT's 2026 AI Conference previews responsible, achievable AI advancements toward next chapter. Focuses on practicality amid rapid innovations.
Builds on global AI momentum.