TLDR : LightOn, a French AI company, announces that ENS Paris-Saclay will be the first to use Paradigm Edu, a sovereign and secure solution designed for higher education. The AI platform will integrate with ENS's existing systems to analyze documents, structure meeting minutes, and potentially assist in other areas such as legal monitoring or student support.
LightOn, a French pioneer in generative AI for businesses and public services, announces that ENS Paris-Saclay will be the first institution to integrate Paradigm Edu, its new offer specially designed for higher education and research actors.
As major American platforms continue to raise concerns related to data extraterritoriality (notably via the Cloud Act), ENS opts for a French turnkey, sovereign, and secure solution. A strategic choice made necessary by the sensitive nature of the data subject to the PPST (Protection of the Nation's Scientific and Technical Potential) device that it may be required to process.
Launched in 2023 by LightOn, Paradigm relies on RAG (Retrieval Augmented Generation), and is regularly updated with new LLMs. The company advocates a frugal approach, aiming to reduce environmental impact, control costs, and accelerate return on investment.
ENS Paris-Saclay is banking on it to accelerate the modernization of all its administrative departments: HR, finance, assets, international relations and research support, communication, information systems...
Paradigm Edu will interface with the institution's internal repositories to produce summaries, analyze strategic documents, and automatically structure meeting minutes. Other use cases are under study, such as helping to respond to project calls, data analysis, legal monitoring, and in a second phase, dedicated assistants to support students in their procedures.
Sylvain Pronteau, Deputy Director of Steering, Quality, and Projects at ENS Paris-Saclay, concludes:
"By choosing Paradigm, we found a reliable technological solution, sovereign, compatible with our internal systems, and integrating environmental issues. This collaboration allows us to concretely explore what generative AI can bring to the public service, while maintaining control over our data and usage"