GRAIL project aims to study and develop novel, general-purpose, generative argumentative AI models, applicable across diverse application domains. By combining advances in natural language processing, machine learning, and computational argumentation, the project aims to develop intelligent agents that can learn, revise, and communicate knowledge through structured arguments and counter-arguments.
The goal is to create a new generation of autonomous hybrid AI systems that learn to reason using (human-like) argumentative reasoning, unlike classical large language models that rely solely on statistical correlations and simply mimic reasoning.
GRAIL agents will autonomously learn argument-based reasoning from text and user interaction, generate new knowledge, update their beliefs based on compelling user input, and build and evolve their own code.
The project proposes a neuro-symbolic AI approach, enabling autonomous agents to interact naturally with humans, learn from dialogue with humans (and other agents), and justify their claims.
Generative AI combined with symbolic AI and agent technology is expected to drive major global economic growth thus GRAIL is well-positioned to generate rapid economic impact, positioning GRAIL as a key French player in the global AI market.
This project is funded by the French National Research Agency (ANR grant ANR-25-CE23-5597).
Members
LIPADE
- Elise Bonzon
- Jérôme Delobelle
- Pavlos Moraitis
- Julien Rossit
- Onn Shehory (Bar-Ilan University)
