AI Safety · NIT Agartala · Open Globally

A research community
for AI Safety.

Weekly paper reading on multi-agent safety, alignment, and AI governance. Every Saturday. Free. Open to students and researchers worldwide.

Backed by
AI Safety India
Parent org · curriculum · mentors
Pathfinder Fellowship
Kairos · Open Philanthropy funded
What We Do

We read. We discuss.
We build toward research.

Every Saturday, we meet to read and discuss AI Safety papers. Topics cover multi-agent safety, alignment and control, interpretability, and governance. No prior knowledge needed to join.

Right now we are a reading club. By mid-2026, members move into a research working group — working on open problems and producing papers, preprints, and policy memos.

Weekly · Saturdays · Free · Online + NIT Agartala
Phase 01 Active now
Paper Reading Club

Structured weekly discussions on core AI Safety papers. Every Saturday. Beginner-friendly. Open to all.

Phase 02 Mid-2026
Research Working Group

Members from Phase 1 move into active research. Open problems. Real output — papers, preprints, policy memos.

Research Focus

Multi-agent safety is our primary focus.

When multiple AI systems interact and coordinate, new failure modes emerge that single-model alignment techniques do not address. This is one of the least studied problems in AI Safety. It is where we spend most of our time.

We also cover alignment and control, interpretability, and AI governance — because these problems connect.

01 — Primary focus

Multi-Agent Safety

Coordination failures, unsafe emergent behavior, and runtime oversight in networks of AI agents.

02

Alignment & Control

Scalable oversight, control evaluations, and deceptive alignment.

03

Interpretability

Mechanistic interpretability, probing, and representation analysis.

04

AI Governance

Policy analysis, risk frameworks, and regulatory research.

Who This Is For

Who we are building this with.

Students

CS, ECE, Data Science. Any year, any university. No prior AI Safety knowledge needed.

Early Researchers

Doing ML work and want to move toward safety-relevant problems.

Experienced Researchers

Already in AI Safety or adjacent fields. Join as a collaborator or contributor.

Mentors

Have experience in technical or governance AI Safety. Help us build from India.

Funders

We are building a research community in India that feeds the global AI Safety field. If that matters to you, reach out.

Open globally

Join AIS NITA →
Founded By

Aditya Raj

Founder, AI Safety India · NIT Agartala
SPAR Fellow Jailbreak Hackathon Top 30 — Grayswan BlueDot Impact Alumni Pathfinder Fellow — Kairos / Open Philanthropy

Aditya started AI Safety India to build the field-building pipeline India does not have. AIS NITA is the research node of that effort — anchored at NIT Agartala, open to the world.

If you want to work on AI Safety and do not know where to start, this is the place.

Expression of Interest

Join AIS NITA

Students, researchers, mentors, and funders — fill in below. We will add you to the community and reach out with event invitations, research program updates, and relevant AI Safety resources.

Join via our Airtable form. Takes 2 minutes.

Join AIS NITA →

Opens in a new tab · Takes 2 minutes

Or join the parent community

Apply to AISIN →
Join AIS NITA →