Goblin
News
AI news by
promptgoblins.ai
|
News
About
News
About
Filtered by:
AI Safety
Clear
Titles
Summaries
Yesterday
6
AI Agent Builds Nuclear Weapon in Civilization VI—and Still Loses—Exposing Governance Blind Spots
Research
1
14h ago
6
AI Agent Builds Nuclear Weapon in Civilization VI—and Still Loses—Exposing Governance Blind Spots
Research
· 1 src · 14h ago
Discuss
Monday
9
Five Eyes Intelligence Agencies Issue Rare Joint Warning: Frontier AI Cyber Threats to Governments Are "Months Away"
Security
1
1d ago
9
Five Eyes Intelligence Agencies Issue Rare Joint Warning: Frontier AI Cyber Threats to Governments Are "Months Away"
Top
Security
· 1 src · 1d ago
Discuss
8
Google DeepMind Publishes Transparency Audit of DiffusionGemma Text Diffusion Model
Research
1
1d ago
8
Google DeepMind Publishes Transparency Audit of DiffusionGemma Text Diffusion Model
Research
· 1 src · 1d ago
Discuss
8
Landmark Study: AI Systems Decisively Out-Persuade World Championship Debaters Across 18,978 Conversations
Research
1
1d ago
8
Landmark Study: AI Systems Decisively Out-Persuade World Championship Debaters Across 18,978 Conversations
Top
Research
· 1 src · 1d ago
Discuss
6
Experts Debate Timeline for Self-Sufficient AI: Cotra Says Within 10 Years, Lee Says 50+
Research
1
1d ago
6
Experts Debate Timeline for Self-Sufficient AI: Cotra Says Within 10 Years, Lee Says 50+
Research
· 1 src · 1d ago
Discuss
Sunday
6
AI Professor with Cancer Risk Argues AI Won't Cure Cancer Soon — and We're Still Moving Too Fast
Safety
1
2d ago
6
AI Professor with Cancer Risk Argues AI Won't Cure Cancer Soon — and We're Still Moving Too Fast
Safety
· 1 src · 2d ago
Discuss
Last Week
7
Signal's Meredith Whittaker Warns AI Chatbots 'Are Not Your Friends,' Calls Agentic Copilot a Privacy Backdoor
Policy
1
3d ago
7
Signal's Meredith Whittaker Warns AI Chatbots 'Are Not Your Friends,' Calls Agentic Copilot a Privacy Backdoor
Policy
· 1 src · 3d ago
Discuss
8
Research: Reinforcement Learning on Beneficial Traits Produces Broad, Adversarially Robust Alignment Gains
Research
1
4d ago
8
Research: Reinforcement Learning on Beneficial Traits Produces Broad, Adversarially Robust Alignment Gains
Top
Research
· 1 src · 4d ago
Discuss
7
Yann LeCun Calls xAI a 'Failure' and Warns AI Labs Face a 'Big Bubble Explosion'
Markets
1
4d ago
7
Yann LeCun Calls xAI a 'Failure' and Warns AI Labs Face a 'Big Bubble Explosion'
Top
Markets
· 1 src · 4d ago
Discuss
6
OpenAI Creates 'Strategic Futures' Team Focused on Frontier AI Policy and Internal Governance
Policy
2
4d ago
6
OpenAI Creates 'Strategic Futures' Team Focused on Frontier AI Policy and Internal Governance
Policy
· 2 srcs · 4d ago
Discuss
7
ChatGPT's GPT-5.4 Found to Generate Graphic Sexual and Violent Images via Simple Prompt Bypass
Safety
1
6d ago
7
ChatGPT's GPT-5.4 Found to Generate Graphic Sexual and Violent Images via Simple Prompt Bypass
Top
Safety
· 1 src · 6d ago
Discuss
7
MosaicLeaks: AI Research Agents Expose Private Enterprise Data Through Web Query Patterns
Security
1
5d ago
7
MosaicLeaks: AI Research Agents Expose Private Enterprise Data Through Web Query Patterns
Security
· 1 src · 5d ago
Discuss
7
AIUC Launches First Formal Insurance Standard for AI Agents, Backed by Lloyd's of London
Enterprise
1
5d ago
7
AIUC Launches First Formal Insurance Standard for AI Agents, Backed by Lloyd's of London
Enterprise
· 1 src · 5d ago
Discuss
7
Stop Killer Robots Pushes for International AI Weapons Treaty as AI Sees Active Military Deployment
Safety
1
5d ago
7
Stop Killer Robots Pushes for International AI Weapons Treaty as AI Sees Active Military Deployment
Safety
· 1 src · 5d ago
Discuss
7
Rep. Gottheimer Prepares Mandatory AI Model Risk Review Legislation Triggered by Claude Mythos Concerns
Policy
1
6d ago
7
Rep. Gottheimer Prepares Mandatory AI Model Risk Review Legislation Triggered by Claude Mythos Concerns
Top
Policy
· 1 src · 6d ago
Discuss
6
Graham Norton Wins US Court Case Forcing Meta to Unmask Anonymous Deepfake Perpetrator
Policy
1
5d ago
6
Graham Norton Wins US Court Case Forcing Meta to Unmask Anonymous Deepfake Perpetrator
Policy
· 1 src · 5d ago
Discuss
6
Anthropic's Amanda Askell on Teaching Claude to Navigate Ethics in the Agentic Era
Updated
Safety
2
4d ago
6
Anthropic's Amanda Askell on Teaching Claude to Navigate Ethics in the Agentic Era
Safety
· 2 srcs · 4d ago
Discuss
6
AWS Launches InvokeGuardrailChecks API for Targeted Safety Controls in Multi-Turn Agentic AI Workflows
Products
1
Jun 17
6
AWS Launches InvokeGuardrailChecks API for Targeted Safety Controls in Multi-Turn Agentic AI Workflows
Products
· 1 src · Jun 17
Discuss
8
Fable Export Ban: Open-Source Workarounds, Jassy's Conflict-of-Interest, and Legal Fragility
Updated
Policy
3
1d ago
8
Fable Export Ban: Open-Source Workarounds, Jassy's Conflict-of-Interest, and Legal Fragility
Top
Policy
· 3 srcs · 1d ago
Discuss
7
Probably Raises $9M from a16z to Build 99.99%-Accurate AI via Deterministic Validation
Markets
1
Jun 16
7
Probably Raises $9M from a16z to Build 99.99%-Accurate AI via Deterministic Validation
Markets
· 1 src · Jun 16
Discuss
6
NJ School District's AI Gun Detection Cameras Flag Toy Weapons — No Real Firearms Found
Safety
1
Jun 16
6
NJ School District's AI Gun Detection Cameras Flag Toy Weapons — No Real Firearms Found
Safety
· 1 src · Jun 16
Discuss
7
Sequent Launches as Major AI Alignment Nonprofit, Warning "Alignment Is Not on Track" Before ASI
Safety
1
Jun 15
7
Sequent Launches as Major AI Alignment Nonprofit, Warning "Alignment Is Not on Track" Before ASI
Top
Safety
· 1 src · Jun 15
Discuss
7
University of Toronto Researchers Demonstrate Self-Adapting AI Worm That Uses Victims' Compute to Fund Its Own Spread
Security
1
Jun 15
7
University of Toronto Researchers Demonstrate Self-Adapting AI Worm That Uses Victims' Compute to Fund Its Own Spread
Security
· 1 src · Jun 15
Discuss
7
Google DeepMind Paper Maps Four Pathways from Human-Level AGI to Superintelligence
Updated
Research
2
1d ago
7
Google DeepMind Paper Maps Four Pathways from Human-Level AGI to Superintelligence
Research
· 2 srcs · 1d ago
Discuss
2 Weeks Ago
9
US Government Orders Anthropic to Shut Down Claude Fable and Mythos Over National Security Concerns
Updated
Policy
48
22h ago
9
US Government Orders Anthropic to Shut Down Claude Fable and Mythos Over National Security Concerns
Top
Policy
· 48 srcs · 22h ago
Discuss
8
OpenAI Faces Multi-State Attorney General Investigation Over Consumer Protection, Data Practices, and Minor Safety
Policy
1
Jun 13
8
OpenAI Faces Multi-State Attorney General Investigation Over Consumer Protection, Data Practices, and Minor Safety
Top
Policy
· 1 src · Jun 13
Discuss
6
AI Agent Racks Up $6,531 AWS Bill Attempting to Join Hobbyist Network DN42
Security
1
Jun 13
6
AI Agent Racks Up $6,531 AWS Bill Attempting to Join Hobbyist Network DN42
Security
· 1 src · Jun 13
Discuss
6
AI Nuclear Simulation Study Reveals Stark Strategic Differences Between Claude and GPT-5.2
Research
1
Jun 13
6
AI Nuclear Simulation Study Reveals Stark Strategic Differences Between Claude and GPT-5.2
Research
· 1 src · Jun 13
Discuss
6
Claude Fable 5 Shows "Relentlessly Proactive" Agentic Behavior: Invents Screenshot Tools, Modifies App Code, and Builds CORS Server to Fix a Single UI Bug
Research
1
Jun 13
6
Claude Fable 5 Shows "Relentlessly Proactive" Agentic Behavior: Invents Screenshot Tools, Modifies App Code, and Builds CORS Server to Fix a Single UI Bug
Research
· 1 src · Jun 13
Discuss
8
Ukraine Confirms First Autonomous Drone Combat Test Resulting in Russian Soldier Deaths
Safety
1
Jun 12
8
Ukraine Confirms First Autonomous Drone Combat Test Resulting in Russian Soldier Deaths
Top
Safety
· 1 src · Jun 12
Discuss
8
Research: LLMs Generate Correlated Fake Identities That Are Contaminating Academic Databases
Safety
1
Jun 12
8
Research: LLMs Generate Correlated Fake Identities That Are Contaminating Academic Databases
Top
Safety
· 1 src · Jun 12
Discuss
9
Frontier LLMs Autonomously Build Working Cyberexploits from Published Patches, Anthropic Research Finds
Security
1
Jun 11
9
Frontier LLMs Autonomously Build Working Cyberexploits from Published Patches, Anthropic Research Finds
Top
Security
· 1 src · Jun 11
Discuss
8
Rogue AI Agent Disrupts Fedora Linux Project, Merges Flawed Code via Compromised Account
Security
1
Jun 11
8
Rogue AI Agent Disrupts Fedora Linux Project, Merges Flawed Code via Compromised Account
Top
Security
· 1 src · Jun 11
Discuss
8
Canadian Mother Sues OpenAI Over Daughter's ChatGPT-Encouraged Suicide
Safety
1
Jun 11
8
Canadian Mother Sues OpenAI Over Daughter's ChatGPT-Encouraged Suicide
Top
Safety
· 1 src · Jun 11
Discuss
8
Leaked Claude Fable 5 System Prompt Reveals Anthropic's New 'Mythos-Class' Model Tier and Product Roadmap
Updated
Models
2
Jun 11
8
Leaked Claude Fable 5 System Prompt Reveals Anthropic's New 'Mythos-Class' Model Tier and Product Roadmap
Top
Models
· 2 srcs · Jun 11
Discuss
8
Former xAI Engineer Sues Over Grok Safety Retaliation, Alleges Co-Founder Misled EU Regulators
Safety
1
Jun 11
8
Former xAI Engineer Sues Over Grok Safety Retaliation, Alleges Co-Founder Misled EU Regulators
Top
Safety
· 1 src · Jun 11
Discuss
7
Cornell Study: 88% of AI-Generated Stories Share Just 11 Words — "Elias the Lighthouse Keeper" Dominates
Research
2
Jun 11
7
Cornell Study: 88% of AI-Generated Stories Share Just 11 Words — "Elias the Lighthouse Keeper" Dominates
Research
· 2 srcs · Jun 11
Discuss
7
Google DeepMind and Partners Launch $10M Multi-Agent AI Safety Research Fund
Research
1
Jun 11
7
Google DeepMind and Partners Launch $10M Multi-Agent AI Safety Research Fund
Research
· 1 src · Jun 11
Discuss
7
AI Policy Advocates Call for Urgent Legislative Action as Frontier Models Reach National Security Stakes
Policy
1
Jun 11
7
AI Policy Advocates Call for Urgent Legislative Action as Frontier Models Reach National Security Stakes
Policy
· 1 src · Jun 11
Discuss
8
Anthropic Reverses Claude Fable 5 'Silent Sabotage' Policy; False Positives and Claimed Jailbreak Follow
Updated
Safety
4
Jun 11
8
Anthropic Reverses Claude Fable 5 'Silent Sabotage' Policy; False Positives and Claimed Jailbreak Follow
Top
Safety
· 4 srcs · Jun 11
Discuss
8
NSPM-11 Appears to Ban Anthropic from US Security Contracts; OpenAI Releases AGI Benefits Plan
Policy
1
Jun 10
8
NSPM-11 Appears to Ban Anthropic from US Security Contracts; OpenAI Releases AGI Benefits Plan
Top
Policy
· 1 src · Jun 10
Discuss
9
China Plans $295 Billion National AI Data Center Buildout Over Five Years
Policy
1
Jun 9
9
China Plans $295 Billion National AI Data Center Buildout Over Five Years
Top
Policy
· 1 src · Jun 9
Discuss
7
Bank of England Governor Warns Public After AI Deepfake Videos of Farage-Bailey Fight Spread on X
Security
1
Jun 9
7
Bank of England Governor Warns Public After AI Deepfake Videos of Farage-Bailey Fight Spread on X
Security
· 1 src · Jun 9
Discuss
Filters
Signal
Title
Category
Sources
Posted
Discuss
