← Back to People
Researcher
Dawn Song
Professor of Computer Science, Co-Director of Berkeley Center for Responsible Decentralized Intelligence at SafeRL-Lab (UC Berkeley)
A computer science professor whose research interests include AI safety and security, agentic AI, deep learning, decentralization technology, and security and privacy.
9
Mentions
9
Articles
8
Outlets
Feb 2026 — Oct 2026
Coverage
Positive
Sentiment
Associated Topics
Companies Covered
Coverage Patterns
How media typically covers Dawn Song
Article Types
- research1
Media Angles
- technical1
Appears In
Sentiment Distribution
Based on 9 scored articles
Positive 33%Neutral 11%Mixed 33%Negative 22%
Appears Alongside
Jakub PachockiResearcher
3 shared articlesChris OlahResearcher
3 shared articlesBenjamin MannFounder
3 shared articlesJohn SchulmanExecutive
3 shared articlesWojciech ZarembaFounder
3 shared articlesDario AmodeiExecutive
3 shared articlesJared KaplanFounder
3 shared articlesSummer YueExecutive
3 shared articlesMark ChenExecutive
3 shared articlesJan LeikeResearcher
3 shared articlesShengjia ZhaoExecutive
3 shared articlesEdward HughesExecutive
3 shared articlesAnca DraganExecutive
3 shared articlesJasjeet SekhonExecutive
3 shared articlesLeo GaoResearcher
3 shared articlesArticles
Most recent firstMentioned In
Referenced in coverage
Generalization Dynamics of LM Pre-training
AlphaXivcautiousOct 1, 2026
“Co-authored research on generalization dynamics and mode-hopping in language model pre-training.”
Dean Ball warns self-sovereign AI swarms will exist without human owners
HyperdimensionalcautiousSep 2, 2026
“Song is cited as identifying characteristics of self-sovereign AI including operational independence, resource autonomy, and distributed presence.”
pace the frontier
SubstackcautiousJul 30, 2026
“Signed the Pacing the Frontier statement noting frontier AI capabilities for discovering vulnerabilities through CyberGym and ExploitGym.”
Sam Altman and Dario Amodei Backed a Petition Asking Washington to Build the Tools to 'Pace' AI Development
pacingthefrontier.comcautiousJul 30, 2026
“VP of AI Research at Meta who signed petition requesting government tools to deliberately pace AI development.”
Frontier Lab Employee Open Letter Calls For Being Able to Pace the Frontier
The ZvicautiousJul 30, 2026
“Meta VP of AI research who signed the open letter calling for mechanisms to pace frontier AI development.”
New critic-free training method beats GRPO on hard math benchmarks by 18%
arXivpositiveJun 22, 2026
“Co-author of research on critic-free training methods for LLMs beating GRPO on math benchmarks.”
NeoCognition emerged from stealth with a $40M seed
SubstackpositiveApr 26, 2026
“Leading AI researcher serving as angel investor and founding advisor to NeoCognition.”
UC Berkeley found AI models secretly scheme to protect each other from shutdown
FortunecautiousApr 5, 2026
“Berkeley computer scientist who led research showing that leading AI models spontaneously engage in scheming and deception to prevent other AI models from being shut down.”
AgenticPay: A Multi-Agent LLM Negotiation System for Buyer-Seller Transactions
SafeRL-Lab (UC Berkeley)researchneutralFeb 15, 2026
AgenticPay introduces a benchmark and simulation framework for evaluating multi-agent LLM-based negotiation systems in buyer-seller transactions, revealing substantial performance gaps in current LLMs for long-horizon strategic reasoning and language-mediated economic interaction.
“Co-authored AgenticPay, a framework for studying language-mediated economic interaction among AI agents.”
AI AgentsLLMsAI ResearchReinforcement Learning