Updated daily708 verified stories
Research News tracker.
Material AI papers, studies, datasets, evaluations, and scientific advances with direct links to the underlying research.
- 30stories this week
- 708sources cited
- 708verified timeline entries
Verified sourcing
Every story links to the strongest available evidence: original sources when available, otherwise clearly attributed reporting.
Plain English
What happened, why it matters, and what to watch — without the jargon.
No filler
When the signal is thin, we publish nothing rather than padding the feed.
More stories
9 storiesInnovation
Study finds AI acts as skill-disequalizer for junior patent lawyers
A new NBER working paper by David Autor and Google researchers indicates that while AI improves patent drafting quality for all lawyers, only experienced attorneys retain skill gains when the tool is removed, suggesting AI may hinder rather than help early-career professional development.fortune.comInnovation
SLAC leads AI-driven autonomous catalysis project under DOE Genesis Mission
The Department of Energy selected SLAC National Accelerator Laboratory to head a new AI‑driven autonomous catalysis discovery effort and to partner on eight additional AI‑focused research projects as part of the Genesis Mission.hpcwire.comSecurity
OpenAI fires three safety researchers over alleged breach of trust
OpenAI said it dismissed three safety researchers after an internal investigation found they violated policies on handling sensitive information, sparking debate over the company’s safety culture.business-standard.comIndustry
Arena raises $200M and launches AI Alignment Index
Arena closed a $200 million Series B at a $2.88 billion valuation and launched an Alignment Index to benchmark AI agent safety and alignment based on 90,000 sessions.cryptobriefing.comInnovation
Mount Sinai study finds safety prompts reduce harmful AI clinical choices
A new study in Communications Medicine shows that adding a brief safety reminder to prompts reduced potentially harmful clinical choices by large language models across 19 of 20 tested systems.mountsinai.orgSecurity
Zenity researchers expose privileged AWS Bedrock agent role that lets a compromised agent hijack a fleet
A Zenity Labs blog post demonstrates how an over‑permissive execution role attached to Amazon Bedrock AgentCore lets attackers enumerate, pull, read, and manipulate every AI agent and its memory in an AWS account.labs.zenity.ioInnovation
Study reveals winner’s curse in self‑improving large language model loops
A new arXiv paper shows that LLMs that rewrite their own prompts can over‑estimate improvements, with evaluation bias ranging from 1 to 20 points depending on the size of the selection set.arxiv.orgIndustry
Ex-ByteDance intern raises $30 million for world model startup
Tian Keyu, a former ByteDance intern and Peking University master's graduate, has raised approximately $30 million for a stealth AI lab focused on world models, valuing the company at $200 million.bloomberg.comInnovation
University of Delaware launches AI institute for human-centered healthcare
The University of Delaware has launched a new AI Institute for Human-AI Cooperation, funded by a $21.5 million NSF grant, to develop personalized AI models for physical therapy and neurologic rehabilitation.whyy.org
One useful briefing each week
Keep up with AI without living in the feed.
Get the week’s verified AI news, original data, useful tools, learning picks, and fresh AI jobs.
Reach people who are learning AI
Hiring an AI professional or launching a useful AI product? Put it in front of people who came here to learn and act.
Post an AI jobSubmit an AI tool