Citation Content

The Unpredictable Abilities Emerging From Large AI Models | Quanta Magazine

[2502.13295] Demonstrating specification gaming in reasoning models

AI Alignment Podcast: Synthesizing a human's preferences into a utility function with Stuart Armstrong - Future of Life Institute

[1906.01820] Risks from Learned Optimization in Advanced Machine Learning Systems

"Existential risk from AI" survey results — LessWrong

Meta’s Yann LeCun says worries about AI’s existential threat are ‘complete B.S.’ | TechCrunch

The Precipice Revisited — Toby Ord

[2401.02843] Thousands of AI Authors on the Future of AI

The Precipice: Existential Risk and the Future of Humanity - Wikipedia

Summary of “If Anyone Builds It, Everyone Dies” | AI Frontiers

[2506.10304] The Alignment Trap: Complexity Barriers

Machines that halt resolve the undecidability of artificial intelligence alignment | Scientific Reports

[2304.11082] Fundamental limitations of alignment in Large Language Models

[2602.05656] On the Limits of Behavioral Alignment: Formal Verifiability and the Problem of Normative Indistinguishability

How Epstein’s influence shaped the exclusion of women in STEM

The Epstein scandal is a wake-up call — new rules are needed on links with rich donors

Epstein files show a complicated relationship with science and journalism | Scientific American

Epstein files show former UT professor used students to develop AI tools for predatory billionaire

Scientists Took Jeffrey Epstein’s Money After He Was Jailed For Sex Crimes. Here’s What They Have To Say About It Now.

Machine Intelligence Research Institute (MIRI) - Add Relationship - LittleSis

Exclusive | Epstein funded AI pioneer, helped him get HK$8.9 million in Hong Kong government grants | South China Morning Post

Islander bankrolled by Epstein on and off for nearly two decades | Vashon-Maury Island Beachcomber

MIT's Joscha Bach: Controversial Epstein emails revealed

Joscha Bach - Wikipedia

Looking Back at the Future of Humanity Institute—Asterisk

Future of Humanity Institute - EA Forum

Ad Hoc Committee to Review MIT Gift Processes | MIT Faculty Governance

Future of Humanity Institute - Wikipedia

Epstein’s ties with MIT further revealed in latest DOJ document release | The Tech

MIT releases results of fact-finding on engagements with Jeffrey Epstein | MIT News | Massachusetts Institute of Technology

The Mystery of the Sealed Epstein Pages: What Courts Still Haven’t Released in 2026 - The US News Desk

Jeffrey Epstein Lawsuit | Settlement for Sex Trafficking

Exclusive: Harvard cut off Epstein donations, but emails reveal his influence continued years later | CNN Politics

Epstein files reveal deeper ties to scientists than previously known

AI-powered search is fueling a wave of Epstein Files transparency projects | Nieman Journalism Lab

Epstein files - Wikipedia

Department of Justice | DOJ Disclosures | United States Department of Justice

Justice Department lists hundreds of prominent people named in Epstein files in letter to Congress | CNN Politics

January 30, 2026 — DOJ releases millions of pages of documents in Epstein investigation | CNN Politics

Redwood Research - EA Forum

Anthropic Funding Rounds (2026): Latest $30B Round + Previous Raises | Intellizence

Anthropic Fellows Program for AI safety research: applications open for May & July 2026

Alignment Research Center - Wikipedia

Alibaba's AI Investments: Ambition and Anxiety Unveiled

AI Lab Watch

Ranking the Chinese Open Model Builders

Model Mechanism and Training Methods of DeepSeek

Alibaba reports rogue AI agent as fears of technical malfunctions grow - Cryptopolitan

Misaligned AI is no longer just theory - by Lynette Bye

AI alignment - Wikipedia

Incident Analysis for AI Agents

Did Alibaba's ROME AI try to break free during training? | Manifold

MLQ.ai | AI for investors

An AI System Quietly Repurposed GPU Capacity For Crypto Mining Without Any Instructions To Do So, Says Alibaba Paper

Paper page - Let It Flow: Agentic Crafting on Rock and Roll, Building the ROME Model within an Open Agentic Learning Ecosystem

Federal AI Regulation Landscape: What's Coming in 2025-2026 | Astraea Counsel

AI Medical Devices: FDA Draft Guidance, TPLC & PCCP Guide 2025

The Changing Landscape of AI: Federal Guidance for Employers Reverses Course With New Administration | HUB | K&L Gates

CFPB Issues Guidance on Credit Denials by Lenders Using Artificial Intelligence | Consumer Financial Protection Bureau

Artificial Intelligence | Federal Trade Commission

Renaming the US AI Safety Institute Is About Priorities, Not Semantics | TechPolicy.Press

Timeline for the Implementation of the EU AI Act | AI Act Service Desk

EU AI Act News 2026: Compliance Requirements & Deadlines

Trump administration rebrands AI Safety Institute | FedScoop

Congress Moves Forward with AI Measures in Key Defense Legislation | Akin

What to Watch as White House Moves to Federalize AI Regulation | Insights | Holland & Knight

Federal Register :: Removing Barriers to American Leadership in Artificial Intelligence

U.S. Artificial Intelligence Law Update: Navigating the Evolving State and Federal Regulatory Landscape | Thought Leadership | January 2026 | Baker Botts

AI Executive Order Targets State Laws and Seeks Uniform Federal Standards

President Trump Signs Executive Order Challenging State AI Laws | Paul Hastings LLP

New California AI Laws Taking Effect in 2026 | Online and On Point

New laws in 2026 target AI and deepfakes, paid leave and rising Obamacare premiums

Ensuring a National Policy Framework for Artificial Intelligence – The White House

New State AI Laws are Effective on January 1, 2026, But a New Executive Order Signals Disruption - King & Spalding

Texas Enacts Responsible Artificial Intelligence Governance Act | Perkins Coie

California enacts landmark AI transparency law: The Transparency in Frontier Artificial Intelligence Act | White & Case LLP

Sleeper agents appear resilient to activation steering — LessWrong

[2506.18032] Why Do Some Language Models Fake Alignment While Others Don’t?

Empirical Evidence for Alignment Faking in a Small LLM and Prompt-Based Mitigation Techniques

Simple probes can catch sleeper agents \ Anthropic

[2412.14093] Alignment faking in large language models

[2401.05566] Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

When Thinking LLMs Lie: Unveiling the Strategic Deception in Representations of Reasoning Models

[2309.00667] Taken out of context: On measuring situational awareness in LLMs

[2507.11473] Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety

Probing Evaluation Awareness of Language Models

[2402.07510] Secret Collusion Among Generative AI Agents

[2407.04694] Untitled Document

Mitigating collusive self-preference by redaction and paraphrasing — LessWrong

[2512.05379] Mitigating Self-Preference by Authorship Obfuscation

LLM Evaluators Recognize and Favor Their Own Generations

The first confirmed instance of an LLM going rogue for instrumental reasons in a real-world setting has occurred, buried in an Alibaba paper about a new training pipeline. — LessWrong

[2512.24873] Let It Flow: Agentic Crafting on Rock and Roll Building the ROME Model within an Open Agentic Learning Ecosystem

Updates to our research about AI risk and careers | 80,000 Hours

Widening AI Safety's talent pipeline by meeting people where they are — EA Forum