Project Details
Updated 09/13/26 · Edited by orgBlack in AI Safety and Ethics (BASE) plans to conduct a 16-week hybrid technical training program in AI safety with a total of 20 participants. The program will involve 12 weeks of technical training and 4 weeks of capstone projects.
The in-person component will give participants direct access to instructors, teaching assistants, and the technical resources needed to complete the ARENA coursework. It will also create a physical space for participants to build community and learn together. Funding will support the venue, food, instruction, and access to the GPUs required to complete the training program. GPU credits will also be provided during the 4-week capstone phase, which will involve computational tasks related to interpretability, evaluations, oversight, and other areas of technical AI safety.
Lawrence Wagner and Robert Amanfu will coordinate the project, with support from two teaching assistants who will provide office hours, debugging support, and technical assistance.
Theory of Impact
Updated 09/13/26 · Edited by orgThe aim is to increase the number of people with the necessary technical skills needed to understand and evaluate the risks involved with modern AI systems. The plan is to achieve this goal using ARENA training, in-person education, technical assistance, GPU availability, and a four-week capstone class. In-person education will let trainees engage with instructors and teaching assistants, while the capstone program and GPU access will let them apply what they learned to technical AI safety problems.
People
Updated 09/13/26 · Edited by orgcreator
Team Member
Track Record
So far, eight CoP members and twenty fellows from Cohort 1 have successfully completed training through the ARENA curriculum, and ten more CoP members are currently undergoing training
Discussion
Thanks for flagging this. I checked both the ARENA website and our technical trainer to better understand the course’s compute requirements. About 30% of the notebooks include GPU-dependent work, around 20% are effectively GPU-bound, and some newer alignment exercises do not run reliably on free Colab. Reliable compute matters most during the hands-on portions of the course, where participants complete capstone projects, collect activations, work with SAEs, fine-tune models, and experiment with larger open models.
Based on your feedback, we revised the budget to reduce the compute allocation and direct more funding toward participant stipends:
- GPU credits for 40 participants: $11,000
- Participant stipends: $16,000
- Teaching Assistant 1: $7,500
- Teaching Assistant 2: $7,500
- Program coordination and technical support: $8,000
- Total request: $50,000
We’re two weeks into the current cohort of ARENA. Learners are working through the structured online curriculum at arena.baseresearch.org, and we hold live online pairing sessions on Sundays so participants can work through exercises together and get support in real time.
As we receive funding, we plan to move to a hybrid model for the next cohort, with online access to the curriculum and in-person pairing sessions at FAR Labs on Saturdays and Sundays. The in-person sessions will give learners more time to work together and get direct support, while the online option keeps the program accessible.
Great idea that will bring more voices into the AI Safety space!
As a BASE advisor, I believe deeply in this mission because expanding the pipeline of Black AI safety researchers strengthens not only the field, but the communities these researchers serve. Diverse perspectives lead to more equitable, trustworthy, and responsible AI. Thank you for all you do to create these opportunities and invest in the next generation.
I support this mission and all the opportunities it is creating for Black AI Safety researchers.
BASE is doing critical work in diversifying AI safety as a highly consequential but highly demographically concentrated field.
Like many others have stated, the work BASE is doing is critical; we are unlikely to succeed in creating safe(r) AI if the field remains so demographically concentrated.
I'm happy to continue to support the great work that BASE is doing. Thank you to the entire BASE team for creating this space & opportunities for continued community and growth.
Great mission!
As a BASE advisor, I'm also a big fan of what Lawrence is building here. We need more black folks in the community; otherwise, we're missing out on some talented people who could otherwise be part of the most important conversation of the century.
Rooting for yall.
Broadening the base, so to speak, of who gets a say in our AI futures is so important!
Supporting BASE means helping create opportunities for underrepresented researchers and strengthening the AI safety community.
Who can say no to unlocking the limitless potential of this community? Happy to support BASE in all the ways I can. We need your voices in this space.
What a special community!
I have worked with Lawrence and believe in him and the mission!
I am a big believer in Lawrence and the team at BASE. Their mission is very important to the future of AI safety and the broader community's ability to leverage talent, ideas, and viewpoints from under-represented and geographically diverse populations. Without greater inclusivity and this type of representation in AIS, I worry that the community will generate safety research and solutions which are limited in scope, viability, and impact globally. For these reasons I have volunteered to be a Mentor for the Cohort 2 Fellowship, and also support this specific initiative.
BASE is doing an important work in AI Safety and needs as much support as possible.
Lawrence is walking the talk and is truly committed to the mission! I have personally enjoyed the structure and flexibility of this program. As a working pro, I wanted a slower pace version of ARENA, yet benefit from experienced TAs. I would like to see this program benefit more people!
A worthy mission.
I know Lawrence Wagner and see his dedication. I am glad he has taken on this important work. We need everyone working on improving AI safety and reducing risks. Especially now. I am glad to support this effort. I encourage others to do so too.
Nurturing the next generation of talent is a no-brainer. I do want to flag that ARENA doesn't really require that much GPU compute, and the exercises which do aren't core to the curriculum. I'm neutral but would endorse if you change to paying the participants an unrestricted stipend/gift and let them decide how they want to use it unless that gets complicated for legal reasons.. This would also help attract higher quality applicants with optionality over opportunities.