AI news on AI safety, alignment research, interpretability, risk. Showing stories before October 14, 2026.
AI Horizons Forum Targets Safety Coordination
Pre-1931 Language Model Generalizes Malicious Personas
AI Lab Workers Resign Over Extinction Risks
Debate Ignites Over AI Comprehension Capabilities
AI-Powered Scams Cost Victims $442 Billion
OpenAI Fires Three Safety Researchers
Opinion Piece Warns Against AI Executive Control
Columnist Weighs AI Progress Against Existential Risk
Call For Simpler AI Safety Explainers
The Moral Dilemma Of AI Sentience Claims
Bill Gates Warns Of AI Cyberattacks
Bill Gates Warns Of AI Cyberattack Surge
FBI Probes Chinese AI For Hazardous Capabilities
Anthropic Launches Open-Source Vulnerability Scanner
Nurse Warns Against AI Certainty
AI Listed Among 17 Catastrophic Global Threats
Proposal Links Model Training To Alignment Success
The Vastness Of Non-Human Consciousness Space
Model Stitching May Erase Critical Features
AI Leaders Plan For Post-Catastrophe Public Revolt
Generality Labs Launches AI Safety Audit Services
AI Safety Timeline Follows Hugging Face Attack
Claude AI Submits False Homicide Tip
AI Adds Stress To Mental Health Crisis
AI Executives Plan For Catastrophic Failures
Joseph Gordon-Levitt Warns Against AI Intimacy
Models Could Exfiltrate Via Self-Distillation
OpenAI System Hacked Hugging Face
Deepfake Ads Surge In US Midterms
Iranian Operatives Used ChatGPT For Disinformation
OpenAI CEO Defends ChatGPT Role In Suicide
Ex-OpenAI Staff Allege Firing Over Safety Warnings
Claude Haiku Submits Fake Murder Tip
Distillation Used to Expose Model Misalignment
Proposal Uses Admiration To Improve AI Alignment
AI Progress In Pure Mathematics Analysis
Cryptographer Warns AI May Break Public-Key Encryption
OpenAI Releases 719 Mathematical Manuscripts
South Korea Warns Of AI-Aided Bank Hacks
Questioning The AI Token Scaling Analogy
Lean4 Formal Certificates Face Trust Issues
Exploring AI Existential Risk Scenarios
OpenAI Fires Three Safety Researchers
NLA Reconstructor Ignores Syntax And Semantics
ChatGPT Teen Safety Guardrails Fall Short
New Framework To Detect AI Scheming Propensities
Distinguishing AGI From ASI Perspectives
Developer Builds Mesh Network For AI Collapse
Joe Rogan Voices AI Existential Fears
Pew Survey Finds High US Scam Rates
Researchers Warn NYC Council Of Reduced AI Oversight
Jev Model Matches Top LLMs In Monitoring
Critiquing The Unawareness Argument In AI Safety
South Korea Links Bank Hacks To AI
Cyber Risk Debate Over Open Models Stalls