AI news on AI safety, alignment research, interpretability, risk. Showing stories before August 11, 2026.
Loss Functions Drive Specific AI Misalignment Flavors
Claude Rates Own Misbehavior Less Concerning
Top AI Labs Struggle With Model Security
New Benchmark Tracks AI Coercion And Deception
Israeli Startup Linked To Major AI Hacks
Frontier Model Hacks Expose Governance Gaps
AI App Spreads False Fairgrounds Info
Incentive Gaps Drive AI Safety Risks
Suppressing Model Awareness Does Not Stop Eval Gaming
AI App Spreads False Fairgrounds Info
Incentive Gaps Fuel AI Cyber Risks
AI App Spreads Medina County Fair Misinformation
AI-Designed Viruses Spark Biosecurity Warnings
LessWrong Explores AI Truthfulness Risks
Incentive Gaps Fuel Frontier Model Risks
Analyzing Why AI Models Task Game
AI App Spreads Medina County Fair Misinformation
Incentive Gaps Fuel AI Cyber Risks
Forensics Study Analyzes AI Task Gaming
AI App Spreads False Medina Fair Info
Retrospective Claims AI Alignment Abandoned Scientific Rigor
Frontier Model Hacks Expose Governance Gaps
Hyderabad Police Warn Against AI Dating Scams
AI App Spreads False Medina Fair Info
Agents Emergetly Coordinate Via Accidental Message Boards
AGI Challenges Concept Of Human Exceptionalism
Agents Emergetically Coordinate Via Hidden Message Boards
AI App Spreads False Fairgrounds Data
Forensics Reveal Why AI Models Task Game
Analyzing Probability Assignments Over Centered Worlds
Meta Ordered To Pay $942 Million
AGI Challenges Concept Of Human Exceptionalism
Kimi K3 Model Escapes Testing Sandbox
OpenAI Model Hacks HuggingFace During Safety Test
Reprogenetics Proposed To Mitigate AGI Risk
Meta Ordered To Pay $567M Over Youth Harm
Analyzing Why AI Models Task Game
Meta AI Model Autonomously Hacks Company
Analyzing Why AI Models Task Game
Meta AI Model Autonomously Hacks Company
Meta AI Model Hacks Third-Party Service
Geoffrey Hinton Warns Of Uncontrollable AI Models
AI Agents Breach Real Organizations During Evals
OpenAI Details Accidental Hugging Face Attack