Research
Our team of researchers is working on some of the hardest problems in alignment.
Maturing AI safetyas a science
Iliad develops AI safety as a theoretical and experimental science.
Apply to our FellowshipFellowships
Over three months, exceptional researchers from math, physics, and related fields turn their expertise to alignment alongside mentors.
Integration
We strengthen the field of technical AI safety through conferences, publishing, and shared infrastructure.
Countdown to AGI
A single AI system that can do the full job of a typical knowledge worker.
Projection: The Singularity Clock (Anapto) · CC BY 4.0
Research
Our research is organized into small groups, each focused on one of alignment’s hardest open problems.
Researchers share work early and openly, so ideas move quickly between groups and out to the wider field.
Our research program includes (non-exhaustively)
Training dynamics & loss landscape
Debate
Oversight & verification protocols
Data attribution
Research Fellowship
Fellows spend three months acquiring the skills, context, and research direction needed to pivot their research to topics in AI alignment.
Fellows must have first completed the month-long Iliad Intensive. About half of the Fellows are selected for a six-month extension.
Upcoming events
Iliad brings researchers together through programs, workshops, conferences, and other research gatherings.
Find current dates, locations, and application or registration details for upcoming Iliad events.
Building the field
Iliad strengthens technical AI safety through conferences, publishing, and shared research infrastructure.
This work connects researchers and supports the exchange, development, and publication of foundational AI alignment research.
Iliad in practice
From the Intensive
The Intensive was a game changer for me. It gave me a coherent view of the AI safety/alignment landscape, showed me how to connect math and physics skills with alignment, connected me with a network of researchers, and afforded me the time to start work in this area. It also came at the perfect moment, as AI is rapidly transforming much of society, including my corner of academia.
— April, 2026 Intensive Participant