Eugene Bagdasarian
@ebagdasa.bsky.social
Assistant Professor at UMass Amherst CICS, Research Scientist at Google Research. https://people.cs.umass.edu/~eugene/
I just gave a keynote at the 18th(!) AI Security workshop CCS'25 in beautiful Taipei. I talked about challenges that future AI agents will face and argued that for defenses we must rely on the context and generate dynamic policies that define what is appropriate to share and do in each context! 1/6
November 6, 2025 at 2:54 PM
I just gave a keynote at the 18th(!) AI Security workshop CCS'25 in beautiful Taipei. I talked about challenges that future AI agents will face and argued that for defenses we must rely on the context and generate dynamic policies that define what is appropriate to share and do in each context! 1/6
Reposted by Eugene Bagdasarian
We're excited to welcome 28 new AI2050 Fellows! This 4th cohort of researchers are pursuing projects that include building AI scientists, designing trustworthy models, and improving biological and medical research, among other areas.
Schmidt Sciences awards $18M to researchers working to ensure AI benefits society - Schmidt Sciences
Media Contact: Carlie Wiener; cwiener@schmidtsciences.org NEW YORK—Schmidt Sciences announced today that 28 scholars studying how to fulfill AI’s potential to dramatically benefit…
www.schmidtsciences.org
November 5, 2025 at 3:43 PM
We're excited to welcome 28 new AI2050 Fellows! This 4th cohort of researchers are pursuing projects that include building AI scientists, designing trustworthy models, and improving biological and medical research, among other areas.
Wanted to get into Multi-Agent System Safety research? Look no more! Terrarium 🐍 by @masonnaka.bsky.social enables different types of attacks: from misalignment to prompt injection to privacy leakage. Realistic tasks, clear metrics, nice abstractions!
🚨New preprint: Terrarium-an open source, blackboard-based testbed for studying safety, privacy & security in LLM multi‑agent systems (MAS). We showcase the vulnerabilities and safety considerations of agentic MASs in this modular and configurable framework. 🧵
#AISafety #LLMAgents #Agents
#AISafety #LLMAgents #Agents
October 31, 2025 at 8:07 PM
Wanted to get into Multi-Agent System Safety research? Look no more! Terrarium 🐍 by @masonnaka.bsky.social enables different types of attacks: from misalignment to prompt injection to privacy leakage. Realistic tasks, clear metrics, nice abstractions!