ILIAD 2: ODYSSEY - Theoretical AI Alignment Conference

ILIAD 2: ODYSSEY is a theoretical AI alignment conference that continues the ILIAD tradition of bringing together researchers working on the mathematical and conceptual foundations of safe and beneficial artificial intelligence. Topics include agent foundations, singular learning theory, scalable oversight mechanisms, causal incentives in AI systems, computational mechanics, and safety-by-debate approaches. Unlike policy-focused AI safety events, ILIAD emphasizes deep theoretical work aimed at understanding the structure of intelligent systems well enough to guarantee their alignment with human values.

The conference is designed for researchers with strong mathematical or philosophical backgrounds who want to contribute to the long-term safety of advanced AI systems. This includes theorists from machine learning, mathematics, philosophy, and cognitive science who are willing to engage with open foundational problems rather than incremental engineering improvements. Graduate students and independent researchers working on alignment theory will find ILIAD a uniquely supportive intellectual environment for sharing early-stage ideas with a knowledgeable and engaged community.

AI alignment AI safety agent foundations singular learning theory scalable oversight causal incentives computational mechanics safety-by-debate