CASI
Join Us

Carnegie
AI Safety Initiative

Working towards a safer future in the presence of powerful artificial intelligence

AI safety is one of the most critical challenges of our time.

CASI brings together researchers, students, and enthusiasts across Carnegie Mellon to advance research, collaboration, and education in AI safety and alignment.

Risks

As artificial intelligence systems become more capable and ubiquitous, ensuring they remain aligned with human values and beneficial to society becomes increasingly important.

{{ risk.num }}

{{ risk.title }}

{{ risk.body }}
Our Mission

Progress on AI safety requires diverse perspectives and interdisciplinary approaches. Our programs meet you wherever you are in your journey.

Past Events

{{ shot.caption }}

Our members have worked with

 {{ org.name }}

Join us as we work together to build a future where AI systems remain beneficial, controllable, and aligned with human flourishing.

Building a safer AI future from Pittsburgh

Our Approach

CASI seeks to spark interest in and facilitate research about AI safety and alignment

{{ risk.num }}

{{ risk.title }}

{{ risk.body }}

What We Offer

Explore the foundations of AI safety through our AI Alignment Fellowship program, which covers technical research challenges, or our AI Governance Fellowship program, which tackles the policy and governance challenges of AI.

Reading Groups

AI Alignment Fellowship

Reading Group Fall/Spring Weekly
Learn More

AI Governance Fellowship

Reading Group Fall/Spring Weekly
Learn More

Research

CASI Research Fellowship

Research Fellowship Spring 2026 Weekly
Learn More

CASI AI Safety Seminar

Seminar Fall/Spring Weekly
Subscribe

CASI Research Scholars

Research Scholars Fall/Spring 10-20 hrs/week
Learn More

Subscribe to our mailing list for updates on our reading groups, research opportunities, events, and everything else happening at CASI.

Join our mailing list

Our team

The exec board leading CASI this year.
{{ p.name }}
{{ p.role }}

Programs We Offer

{{ p.title }}

{{ p.timing }}

{{ p.desc }}

{{ p.descBold }}
{{ p.cta }}

CASI Core Membership

Join our most dedicated community members.

CASI Core Membership recognizes and supports our most active, enthusiastic members who have demonstrated potential to make an impact on AI safety. As CASI interacts with many people across different programs and events, we created core membership to deepen our connections with those who are most committed to our mission.

Applications closed {{ coreLabel }}

For the full terms and norms of the program, see the CASI Core Member Constitution.

Core membership benefits

Community & space

•Bi-weekly invite-only socials at the Nest, with catered food and guests

•24/7 access to the Nest: desks, reading room, stocked snacks

•Regular one-on-ones with core members and exec

Guests & retreats

•Priority access to guest events: dinners with researchers like Stephen Casper (MIT), Nathan Lambert (Ai2) and Nikola Jurkovic (METR)

•Fully-paid AI safety retreats: Action Potential, AISST/MAIA, OASIS

Funding & research support

•Conference and travel funding

•Compute and research expense funding

•Application sprints for MATS and Astra

Growth & perks

•Run your own events with CASI support

•Contribute to existing programming

•Free merch

Frequently Asked Questions

Everything students usually ask before joining CASI. Still unsure? Email us at casi@andrew.cmu.edu.

{{ item.q }}

{{ item.a }}

Get in touch

Reach out about membership, programs, or collaboration.

Stay updated with the latest AI safety research, upcoming events, and community news. Get exclusive access to our research insights and be the first to know about workshops, speaker events, and collaboration opportunities.

Join our mailing list
Email
casi@andrew.cmu.edu

General inquiries, membership questions and collaboration.

Discord
Join the server

Real-time discussion, event updates and community.

Office
Craig-Henry Building

Units 1M and 1E, 201 South Craig Street, Pittsburgh, PA 15213

Events

See what we have planned. Subscribe to our calendar

{{ ev.wd }}
{{ ev.day }}
{{ ev.time }}
{{ ev.place }}
{{ t.label }}

{{ ev.title }}

{{ ev.blurbRich }}
{{ eventPageLabel }}

What we organize

Guest lectures

Leading AI safety researchers share their insights and latest findings with our community.

Workshops

Hands-on sessions covering AI safety techniques, research methodologies, and practical applications.

Discussion groups

Regular meetups to discuss recent papers, research directions, and emerging topics in AI safety.

Networking events

Connect with fellow students, researchers, and faculty interested in AI safety and alignment.

Research presentations

Students and faculty present their ongoing research in AI safety and related fields.

Socials

Enjoy free food and get to know people you otherwise wouldn't meet on campus.

CASI Programming

AI Alignment Fellowship

An introduction to technical AI safety research

An accessible introduction to cutting-edge AI safety research, with weekly paper discussions and food provided.

Our AI Alignment Fellowship runs every semester and provides an accessible introduction to cutting-edge AI safety research. We cover essential topics including RLHF, jailbreaking, interpretability, AI control, and more.

Format

The group meets weekly to discuss carefully selected papers that build foundational knowledge in technical AI safety. Food is provided at every meeting and no work is required outside of our sessions. Just show up ready to engage with the research and the discussion.

Goals

The goal of this reading group is to give you an overview of the topics in technical AI safety. Much of what you see is just the tip of the iceberg. By the end, you should be able to:

•Reason about arguments and timelines for existential risk from AI.

•Ask substantive questions about technical problems in AI safety.

•Brainstorm concrete ideas for problems in technical AI safety.

Curriculum

You can view the materials and papers from our previous iteration to get a sense of the topics we cover and the depth of discussion.

AI Governance Fellowship

Exploring AI governance and policy

An accessible introduction to the policy and governance challenges of advanced AI, with weekly discussions and food provided.

Our AI Governance Fellowship runs every semester and provides a comprehensive introduction to AI governance and policy research. We cover essential topics similar to the Blue Dot curriculum, including AI regulation frameworks, international cooperation on AI safety, risk assessment methodologies, and governance mechanisms for advanced AI systems.

Format

The group meets weekly to discuss policy papers, governance proposals, and real-world case studies that shape how society manages AI development and deployment. Food is provided at every meeting and no work is required outside of our sessions. Just come ready to engage with critical policy questions and thoughtful discussion about AI's societal impact.

Who it is for

Whether you are interested in law, public policy, international relations, or simply want to understand how AI governance works, this reading group offers an accessible entry point into one of the most important policy challenges of our time.

CASI Research Fellowship

Previously known as the Project Group

Hands-on research experience in AI safety for motivated students who are ready to dive deeper into the field.

Our Research Fellowship provides hands-on research experience in AI safety for motivated students ready to dive deeper into the field. This program requires an extended application process and is designed for students who have demonstrated commitment to AI safety research.

Prerequisites

Applicants must have either completed one of our introductory reading groups with good attendance, or demonstrate equivalent experience in AI safety research or related fields.

Mentorship

Spring 2026 is mentored by PhD students Jennifer Evans and Ida Mattson, with guest mentors brought in by project.

CASI Research Scholars

Structured research opportunities with expert mentorship

Pre-scoped projects that pair undergraduate and masters students with professors, postdocs, and senior PhD students.

What is the CASI Research Scholars program?

The CASI Research Scholars program connects talented undergraduate and masters students with professors, postdoctoral scholars, and senior PhD students to work on cutting-edge AI safety research. This program differs from our Research Fellowship in that research projects are pre-scoped by PhD mentors, allowing you to join a well-defined research initiative with clear direction and support.

How is this different from the Research Fellowship?

While our Research Fellowship is exploratory, where you help define your own research direction through team matching and scoping, the Research Scholars program provides pre-scoped projects mentored by PhD students. This allows for deeper focus and more immediate research engagement.

Application process

Each research project has its own application process. Applications typically require around 10 hours to complete, which may include project-specific assignments, essays, or research proposals. This ensures we find mentees who are genuinely excited about the project at hand.

What you will gain

Expert mentorship

Work directly with PhD students and experienced researchers.

Research skills

Develop real research competencies in AI safety.

Publications

Contribute to research that is concretely directed towards a publication.

Compute and API credits

CASI provides the resources you need to conduct your research.

Impact

Contribute meaningfully to the AI safety research community.

Time commitment

10 to 20 hours per week, varying by project, across collaborative and independent work.