"Why did the AI do that?" 🤔 When an LLM misbehaves, the cause is rarely simple. @dr-bly.bsky.social breaks down the 5 complex factors that interact to create surprising AI behavior. Read the full post: cset.georgetown.edu/article/why-...

Why Do AI Systems Misbehave? | Center for Security and Emerging Technology

AI systems are increasingly impressive, which makes their failures all the more baffling. This blog dives into the causes of AI misbehavior.

cset.georgetown.edu

I gave my thoughts to AP's Matt O'Brien about the OpenAI hack of Hugging Face The details are still scant, but there are a few characteristics of this incident that I find interesting 1) This is the highest level of autonomy that we've seen in the use of a large language model for cyber operations

CSET@csetgeorgetown.bsky.social · 2w ago

🚨"This is the highest level of autonomy that we've seen in the use of a large language model for cyber operations," CSET’s @dr-bly.bsky.social told @apnews.com re: the OpenAI models that broke out of their sandboxed testing environment & into Hugging Face’s production servers.🚨 shorturl.at/SKUdD

Check out my latest CSET article on how the components of an AI system impact its behavior. Written for a non-technical audience, this was an exercise in explaining things "as simple as possible, but not simpler."

CSET@csetgeorgetown.bsky.social · 3w ago

Ever ask yourself “Why did the AI do that?!” after it makes a baffling mistake? This new CSET blog has the answer! This is your entry point for a deeper dive into what causes AI systems to misbehave: cset.georgetown.edu/article/why-...

🚨 We're Hiring! 🚨 CSET is looking for the right person to build and lead our Frontier AI team! Ideal candidates bring deep expertise in frontier AI, large-scale model development compute infrastructure, or China's AI policy ecosystem. Apply below! cset.georgetown.edu/job/research...

Frontier AI Research Lead | Center for Security and Emerging Technology

CSET Frontier AI Research Lead The Center for Security and Emerging Technology (CSET) is currently seeking candidates to lead our Frontier AI research efforts, either as a Research Fellow or Senior Fe...

cset.georgetown.edu

Excited to share a new op-ed by @minanrn.bsky.social, @jessicaji.bsky.social, and myself for the National Interest! The administration's new AI Executive Order, aiming to suppress state-level AI regulation, risks undermining the innovation it seeks to advance. nationalinterest.org/blog/techlan...

The Complicated Politics of Trump’s New AI Executive Order

The administration’s attempt to suppress state AI regulation risks legal backlash, bipartisan resistance, and public distrust, undermining the innovation it seeks to advance.  The Trump administration...

nationalinterest.org

Helping to organize and run this workshop has been a highlight of my time at CSET. My takeaways from the workshop are that it's hard to find strong signal on the future of AI use for automating R&D - so we need as much insight into how AI is being used at present for AI R&D.

CSET@csetgeorgetown.bsky.social · 6mo ago

⭐️New CSET Report⭐️ Are we headed toward an "intelligence explosion"? Our new workshop report led by @hlntnr.bsky.social on automating AI R&D digs into how AI companies are using their own AI models to accelerate their research, and what could happen next 👀👇 cset.georgetown.edu/publication/...

My AI Governance and National Policy course is a wrap! I covered the technical background of AI, how AI is applied (and why we might want to govern it), and what policies about AI are out there. Find my syllabus here (and let me know what I'm missing): docs.google.com/document/d/1...

AI Gov & Nat'l Policy '25

Week Theme Week 1 Sep 2 Introduction and AI Background The story of AI so far, and what deep learning is. Extra Resources: Computer History Museum: https://www.computerhistory.org/timeline/ai-roboti...

docs.google.com

Check out my new @csetgeorgetown.bsky.social report, written alongside @minanrn.bsky.social, @jessicaji.bsky.social, and Ngor Luong! cset.georgetown.edu/publication/... Identifying assumptions can help policymakers make informed, flexible decisions about AI under uncertainty.

AI Governance at the Frontier | Center for Security and Emerging Technology

This report presents an analytic approach to help U.S. policymakers deconstruct artificial intelligence governance proposals by identifying their underlying assumptions, which are the foundational ele...

cset.georgetown.edu

Earlier this month I had a great conversation about #AI and #security over lunch with Apolline Rolland. You can read about it here virtual-routes.org/ai-over-lunc... Hopefully you can find some useful insights in my (very well edited) ramblings!

AI over Lunch: Colin Shea-Blymyer

This week’s AI over Lunch comes from Washington DC, where we meet Colin Shea-Blymyer, researcher at the Centre for Security and Emerging Technology (CSET), to talk about AI beyond the European bubble.

virtual-routes.org

This is an awesome opportunity. Come work with awesome researchers (and me, too), tackle the thorniest debates in AI, and make real impact!

CSET@csetgeorgetown.bsky.social · 10mo ago

CSET is hiring 📢 We’re seeking a Research or Senior Fellow to help lead our Frontier AI policy research efforts. Interested or know someone who would be? Learn more and apply 👇 cset.georgetown.edu/job/research...

Yesterday I taught my first class. I'm officially a teacher (a professor, even)! I'm very excited to be teaching AI policy to undergrads at Georgetown this semester.

CSET is hiring! Be sure to apply by Tuesday! Join our data team as a Data Research Analyst to contribute to data-driven research products and policy analysis at the forefront of national security and tech policy. cset.georgetown.edu/job/data-res...

Data Research Analyst | Center for Security and Emerging Technology

We are currently seeking capable data storytellers, analyzers, and visualizers to serve as Data Research Analysts.

cset.georgetown.edu

Banning state-level AI regulation is a bad idea! One crucial reason is that states play a critical role in building AI governance infrastructure. Check out this new op-ed by @jessicaji.bsky.social, myself, and @minanrn.bsky.social on this topic! thehill.com/opinion/tech...

2 weeks left on this open funding call on risks from internal deployments of frontier AI models—submissions are due June 30. Expressions of interest only need to be 1-2 pages, so still time to write one up! Full details: cset.georgetown.edu/wp-content/u...

Helen Toner@hlntnr.bsky.social · last yr.

💡Funding opportunity—share with your AI research networks💡 Internal deployments of frontier AI models are an underexplored source of risk. My program at @csetgeorgetown.bsky.social just opened a call for research ideas—EOIs due Jun 30. Full details ➡️ cset.georgetown.edu/wp-content/u... Summary ⬇️

💡Funding opportunity—share with your AI research networks💡 Internal deployments of frontier AI models are an underexplored source of risk. My program at @csetgeorgetown.bsky.social just opened a call for research ideas—EOIs due Jun 30. Full details ➡️ cset.georgetown.edu/wp-content/u... Summary ⬇️

Pope Leo XIV is a rare Virgo ♍

I've been keeping a list of organizations that do AI red-teaming. Sources are blog and job postings. Definitely an incomplete list. Potentially out of date. I might do a write-up about it. docs.google.com/document/d/1...

AI Red Teams List

This document serves as my place to track and organize AI red teaming activity. Model Developers Organizations in this section are directly involved in the development of models and have expertise in...

docs.google.com

"We all rely on science [...] Businesses and farmers rely on science and engineering for product innovation, technological advances, and weather forecasting. Science helps humanity protect the planet and keeps pollutants and toxins out of our air, water, and food." docs.google.com/document/d/1...

Public Statement on Supporting Science for the Benefit of All Citizens

TO THE AMERICAN PEOPLE We all rely on science. Science gave us the smartphones in our pockets, the navigation systems in our cars, and life-saving medical care. We count on engineers when we drive acr...

docs.google.com

ICYMI: The CSET webinar on AI red-teaming has been recorded! www.youtube.com/watch?v=gDnN... Watch this for a great discussion on what AI red-teaming is, how different organizations do it, and how it can be improved! Huge thanks to the panelists, moderator, and audience!

What’s Next for AI Red-Teaming?

YouTube video by Center for Security and Emerging Technology

youtube.com

Almost everything we care about in AI comes down to evaluations, including #redteaming. Jessica Ji's post lays out a path for making AI red-teaming better. cset.georgetown.edu/article/how-...

How to Improve AI Red-Teaming: Challenges and Recommendations | Center for Security and Emerging Technology

Despite recent upheaval in the AI policy landscape, AI evaluations—including AI red-teaming—will remain fundamental to understanding and governing the usage of AI systems and their impact on society. ...

cset.georgetown.edu

Starting in 5 👀

CSET@csetgeorgetown.bsky.social · last yr.

What: CSET Webinar 📺 When: Tuesday, 3/25 at 12PM ET 📅 What’s next for AI red-teaming? And how do we make it more useful? Join Tori Westerhoff, Christina Liaghati, Marius Hobbhahn, and CSET's @dr-bly.bsky.social * @jessicaji.bsky.social for a great discussion: cset.georgetown.edu/event/whats-...

🏀 I know nothing about college basketball, so I decided to be the avatar for ChatGPT in the @csetgeorgetown.bsky.social #MarchMadness tournament. I used Deep research on o3-mini-high to generate a detailed report and analysis of the tournament. The resulting bracket is very conservative. Go Humans!

✨ NEW: How can the U.S. stay ahead in AI? OSTP called for input on developing an “AI Action Plan.” Our response outlines the steps the U.S. should take to: 1️⃣Secure and advance its AI leadership 2️⃣Navigate competition with China 3️⃣Realize AI’s benefits and avoid its risks

Why does #AI red-teaming suck? How can we make it suck less? All this and more at the next CSET webinar: cset.georgetown.edu/event/whats-... Join me, the Director of Microsoft's AI Red Team, the MITRE ATLAS Lead, and the Director of Apollo Research. This will be an awesome conversation.

What’s Next for AI Red-Teaming? | Center for Security and Emerging Technology

On March 25, CSET will host an in-depth discussion about AI red-teaming — what it is, how it works in practice, and how to make it more useful in the future.

cset.georgetown.edu

What: CSET Webinar 📺 When: Tuesday, 3/25 at 12PM ET 📅 What’s next for AI red-teaming? And how do we make it more useful? Join Tori Westerhoff, Christina Liaghati, Marius Hobbhahn, and CSET's @dr-bly.bsky.social * @jessicaji.bsky.social for a great discussion: cset.georgetown.edu/event/whats-...

What’s Next for AI Red-Teaming? | Center for Security and Emerging Technology

On March 25, CSET will host an in-depth discussion about AI red-teaming — what it is, how it works in practice, and how to make it more useful in the future.

cset.georgetown.edu

I gave my takes from the Paris AI Action Summit to @thecipherbrief.bsky.social last week. I heard three Nobel Laureates warn about the risks of AI, while VP Vance was "not here ... to talk about AI safety" and focused on opportunity. I think we can innovate on AI without building unsafe products.

@miahoffmann.bsky.social , @ojdaniels.bsky.social, and I wrote a piece on key AI governance areas to watch in 2025 with the upcoming AI Action Summit in mind. Check it out here! thebulletin.org/2025/02/will...

Will the Paris artificial intelligence summit set a unified approach to AI governance—or just be another conference?

AI innovations and governments’ preferences can make international consensus on governance at the Paris Summit challenging.

thebulletin.org

We're hiring 📢 CSET is looking for a Research Fellow to analyze topics related to the development, deployment, and operations of AI & ML tools in the national security space. Interested or know someone who would be? Learn more and apply 👇 cset.georgetown.edu/job/research...

Research Fellow - Applications | Center for Security and Emerging Technology

The Center for Security and Emerging Technology at Georgetown University (CSET) is seeking applications for a Research Fellow to support our Applications Line of Research. This role will analyze topic...

cset.georgetown.edu

CSET is hiring 📢 We’re looking for our next Director of Analysis to lead our research agenda & manage CSET's talented team of researchers. Interested or know someone who would be? Learn more and apply 👇 cset.georgetown.edu/job/director...

Director of Analysis | Center for Security and Emerging Technology

The Center for Security and Emerging Technology at Georgetown University (CSET) is seeking applications for a Director of Analysis. This leadership role combines organizational strategy, research over...

cset.georgetown.edu

I'm packing my bags for an exciting week in Paris! I'll be contributing to the Paris Peace Forum's AI-Cyber Nexus discussions, and I'll be attending @iaseai.bsky.social '25 and the Paris AI Security Forum. I'm registered for a few other events, as well (if I can find the time and energy to attend).