<aside> 📣

Spring 2027: We are considering running our second ARENA cohort, fill out our interest form to help us gauge demand!

</aside>

We exist to connect UF students who are passionate about AI safety, and to train them until they can contribute to the field.

Our programs cultivate real technical skills and culminate in access to real research and networking opportunities with professionals working in the field. Following our first cohort, we are currently gauging interest for a second free, in-person cohort in Spring 2027 working through ARENA, the technical curriculum the AI safety field trains on. Tell us where you're at technically and register your interest below to help shape our Spring plans.

Why AI Safety Matters

AI is getting more capable faster than anyone is making it safe. In September 2026, Anthropic's CEO Dario Amodei wrote that "we must slow the pace at which we improve the capabilities of AI models," because left unchecked, progress "could outrun our ability to understand and control these systems," and that "there's much more to do to ensure that our alignment training keeps up with the growth in model capabilities." (We Must Pace the Frontier) What the field is short of is people who can do that technical work. That is what we exist to change.

New to AI safety? Start here.

If you already know why you're here, skip to the cohort. If you're still deciding whether this is real, these are the questions people ask us most.

• Is this a real field, or science fiction?
• What actually happened this year?
• Can I actually contribute?
• What does the work look like day to day?
• Is this a career, or a cause?

AI Safety Upskilling Sprint · Spring 2027

Our sprint relies on ARENA’s hands-on technical curriculum. Following our first run, we are looking to host a second small, in-person peer cohort this Spring: two sprints a week, in pairs. You build GPT-2 from scratch, look inside it, train it with RL, and learn to evaluate it.

<aside> ➡️

Register interest

✍️ Spring 2027 ARENA Interest Form

(Tell us where you're at technically so we can gauge demand and finalize the cohort!)

</aside>

<aside> 📋

At a glance

• Format: in-person peer sprints, twice a week. Pairs inside a group of about five, with a facilitator who has done the material.
• Time: roughly 8 hours a week for the length of the cohort. Exact schedule set at kickoff.
• Requires: linear algebra and the full calculus series, done. Fundamentals of Machine Learning strongly preferred.
• Cost: free.
• Status: We are gauging interest now. Fill out the interest form so we know where you're at and can notify you first when Cohort 2 seats are finalized.
• Seats: one small cohort. If there is enough demand to run it this Spring, spots will be capped to keep the peer pairs tight. </aside>

Meet the math requirements but borderline on the rest—or still working through calculus/ML this semester? Fill out the interest form anyway and tell us where you're at. It helps us tailor our Spring plans, and we can point you toward what to work on in the meantime.

• What you'll build: the four ARENA chapters

Who we look for

The same things the research programs select for. MATS describes its scholars as people "deeply motivated by AI alignment" and "equipped with the skills and mindset to contribute to cutting-edge research." For our prospective Spring cohort, that means three things: