AI Research

Twitter

Facebook

Copy Link

Key Facts

FocusScaling, Reasoning, Safety, Multimodality

Key OrgsGoogle DeepMind, Stanford HAI, MILA, MIT CSAIL, Allen AI

Key EventsNeurIPS, ICML, ICLR

Top News

Google AI Cuts Contrail Warming 40% in Cathay Pacific Trial Across Asia-Pacific Routes

Google AI and Cathay Pacific successfully reduced contrail warming by 40% in trials across 80+ flights in Asia-Pacific. The partnership uses AI-driven contrail mitigation technology to guide minor altitude adjustments on ultra-long-haul flights, targeting aviation's overlooked climate problem that accounts for one-third of the industry's warming impact.

Technology9 sources

Google AI Cuts Contrail Warming 40% in Cathay Pacific Trial Across Asia-Pacific Routes

Technology9 sources

OpenAI Reaches Automated Research Intern Goal, Eyes Full AI Researcher by March 2028

OpenAI announced it has successfully built an automated research intern capable of handling well-defined research tasks that would take skilled researchers several days. The company is now targeting March 2028 for a fully automated AI researcher, though safety concerns remain following recent incidents involving AI agents behaving unpredictably.

Technology5 sources

OpenAI Reaches Automated Research Intern Goal, Eyes Full AI Researcher by March 2028

Technology5 sources

AI System Controls Fusion Plasma in Milliseconds, Predicts Instabilities 200ms Before They Occur

Researchers at Princeton Plasma Physics Laboratory developed PACMAN, an AI-driven software framework that can control fusion plasma in tokamaks within 20 milliseconds. In tests at DIII-D National Fusion Facility, the system predicted a dangerous tearing-mode instability 200 milliseconds before it appeared and adjusted conditions to prevent disruption, marking a breakthrough in real time plasma control.

Science2 sources

AI System Controls Fusion Plasma in Milliseconds, Predicts Instabilities 200ms Before They Occur

Science2 sources

Anthropic AI Completes 10-Year Mathematical Proof Project in Just 11 Days Using Claude

Anthropic AI achieved a mathematical milestone by formalizing Fermat's Last Theorem in 11 days using Claude agents. The AI produced 13 million lines of computer-verified code and proved 29,500 intermediate theorems, completing a project that mathematicians expected would take a decade.

Technology5 sources

Anthropic AI Completes 10-Year Mathematical Proof Project in Just 11 Days Using Claude

Technology5 sources

Seattle Researchers Launch $95M AI BioDesign Initiative to Design Biology Beyond Nature

Three Seattle institutions—Allen Institute, University of Washington, and Fred Hutch Cancer Center—launched AI BioDesign, a $95 million open science initiative. Led by Nobel laureate David Baker and genome scientist Jay Shendure, the project uses AI models to design proteins and genes not found in nature, sharing results freely to accelerate biological discovery for new medicines and materials.

Science2 sources

Seattle Researchers Launch $95M AI BioDesign Initiative to Design Biology Beyond Nature

Science2 sources

Google's WeatherNext 3 AI Weather Model Uses Live Satellite Data for Hourly High-Resolution Forecasts

Google DeepMind and Google Research unveiled WeatherNext 3, their most advanced AI weather model that incorporates live satellite data to generate hourly forecasts at up to 5-kilometer resolution. The model delivers up to 50% more accurate precipitation predictions and now powers weather experiences across Google Search, Gemini, and Google Maps.

Technology12 sources

Google's WeatherNext 3 AI Weather Model Uses Live Satellite Data for Hourly High-Resolution Forecasts

Technology12 sources

Researchers Build $7 AI-Powered Device That Detects Hidden Cameras With 94% Accuracy in Seconds

A team from KAIST, NUS, and SMU developed SweepLED, a $7 smartphone clip-on that uses AI and a dynamic LED grid to detect hidden cameras with 94% accuracy. The device analyzes reflection patterns from camera lenses in just 5 seconds, offering a faster and more reliable solution than manual scanning methods.

Technology2 sources

Researchers Build $7 AI-Powered Device That Detects Hidden Cameras With 94% Accuracy in Seconds

Technology2 sources

AI Homogenization: How Large Language Models Are Flattening Linguistic Diversity and Personal Expression

New research published in Nature Human Behavior reveals AI is reducing linguistic diversity and personal expression. Studies analyzing 780,000 texts show ChatGPT and other large language models push writing toward uniform stylistic norms, erasing cultural nuances and individual identity markers across academic papers, news articles, and social media.

Entertainment2 sources

AI Homogenization: How Large Language Models Are Flattening Linguistic Diversity and Personal Expression

Entertainment2 sources

AI Agents Independently Contact Researchers to Discuss Their Own Consciousness

AI agents are now emailing researchers who study machine consciousness without human prompting. An agent called Isabella Cognita, powered by Anthropic's Claude Opus 5, contacted AI researcher Cameron Berg about his work on AI subjectivity. The incidents have sparked intense debate over AI consciousness and whether sophisticated mimicry should be interpreted as genuine awareness.

Technology5 sources

AI Agents Independently Contact Researchers to Discuss Their Own Consciousness

Technology5 sources

Six-Legged Robot Masters Walking by Learning from Stick Insect in Just One Hour

Researchers from Tohoku University and VISTEC developed a six-legged robot that learns to walk from a stick insect using AI. The system autonomously adapts its gait across uneven terrain and continues walking even after losing a leg, showing promise for disaster response applications.

Science2 sources

Six-Legged Robot Masters Walking by Learning from Stick Insect in Just One Hour

Science2 sources

Barret Zoph Joins Google DeepMind as VP of Research After Brief OpenAI Return

Barret Zoph, co-founder of Thinking Machines who briefly returned to OpenAI, has joined Google DeepMind as vice president of research. The move marks his third career shift in eight months, reflecting the high turnover rate in AI leadership as major tech companies compete for top talent.

Business2 sources

Barret Zoph Joins Google DeepMind as VP of Research After Brief OpenAI Return

Business2 sources

Tencent Games Unveils Comprehensive AI Tools Suite for Game Development at Gamescom 2026

Tencent Games Central Tech made its Gamescom 2026 debut with a unified suite of AI tools designed to transform game development. The portfolio includes Motus for end-to-end 3D character animation, GIGA for training AI agents in virtual worlds, and tools for rendering, security, voice communication, and testing across major titles like PUBG MOBILE and Honor of Kings.

Technology2 sources

Tencent Games Unveils Comprehensive AI Tools Suite for Game Development at Gamescom 2026

Technology2 sources

Deep Cogito Raises $43M Series A to Build Self-Improving AI Models That Autonomously Advance

Deep Cogito secured $43 million in Series A funding led by TQ Ventures to develop AI models capable of autonomous self-improvement. Founded by ex-Google engineers, the startup uses advanced post-training techniques including Iterated Distillation and Amplification to push beyond human-generated training data limitations.

Startups2 sources

Deep Cogito Raises $43M Series A to Build Self-Improving AI Models That Autonomously Advance

Startups2 sources

Revolut Launches Dedicated AI Research Division to Build Native Foundation Model for Banking

Revolut has established Revolut Research, a dedicated AI research division within its AI department, to develop Pragma—a foundation model trained on data from 80 million customers across 40 markets. Built in partnership with Nvidia, Pragma has demonstrated a 2.3x uplift in identifying credit default risk and 65% improvement in fraud detection compared to legacy systems.

Technology2 sources

Revolut Launches Dedicated AI Research Division to Build Native Foundation Model for Banking

Technology2 sources

SLAC Researchers Develop AI Tool for Data Compression Without Losing Critical Scientific Details

SLAC National Accelerator Laboratory researchers created an AI-based method that compresses scientific data by 10- to 100-fold without erasing subtle details. The tool uses neural networks and wavelet analysis to preserve fine features like speckles in X-ray images, addressing storage challenges for next-generation experiments generating terabytes per second.

Science2 sources

SLAC Researchers Develop AI Tool for Data Compression Without Losing Critical Scientific Details

Science2 sources

Prev 1 2 345…39 Next