Tag
New research from Google and UVA reveals that longer AI reasoning traces actually correlate with wrong answers. The fix: measure how deeply the model thinks, not how much it writes.
New paper shows 'intent laundering' bypasses Gemini, Claude, and other models with 90-98% success by removing obvious attack cues
UCSF study finds generative AI can build prediction models in minutes that took human teams months, though only half the tested systems worked.
University of New Hampshire researchers used AI to scan 67,000 compounds and find alternatives to rare earth magnets critical for EVs and clean energy.
Mount Sinai researchers tested 20 LLMs with over a million prompts and found they readily accept false medical claims embedded in clinical-looking documents.
Researchers discovered that displaying an AI model's reasoning process creates a roadmap for attackers. OpenAI's o1 rejection rate dropped from 98% to under 2%.
An Emory study found that pairing clinical staff with AI tools improved accuracy in identifying eligible cancer patients without adding to workload.
Anthropic's research shows that explicitly permitting reward hacking prevents models from generalizing to sabotage and deception
Two separate projects used AI to systematically mine decades of archived telescope data, pulling out hundreds of never-documented cosmic anomalies and over a million variable objects that human review had overlooked.
University of New Hampshire team built an AI that extracted magnetic data from decades of research, identifying 25 new high-temperature magnets that could replace rare earth elements in EVs.
Harvard team's foundation model trained on 49,000 brain MRIs outperforms specialized AI tools at predicting dementia risk, brain cancer survival, and tumor mutations.
New agentic AI system from Shanghai Jiao Tong University correctly identifies rare diseases more accurately than human specialists, tested across 6,401 cases.
New research shows AI reasoning models can autonomously plan and execute attacks that bypass safety guardrails in nearly all other AI systems.
Singapore researchers combine AI with physics simulations to predict protein structures 13% more accurately than existing methods, covering 73% of the human proteome.
OpenScholar matches human expert citation accuracy while GPT-4o fabricates sources 78-90% of the time. The code, models, and 45 million paper corpus are all free to use.
Google's upgraded reasoning model finds flaws in peer-reviewed papers, optimizes semiconductor fabrication, and outperforms every frontier model on scientific benchmarks.