A Secure Generative AI for Businesses: The Promising Initiative by Efrei Students

Students from Efrei have developed a generative artificial intelligence capable of quickly integrating into businesses to improve information management while ensuring data security.

A Student Project with Professional Ambitions

According to the site l'Usine Nouvelle, as part of their final year project, four students from Efrei — Matthieu Evuort, Marty Evuort, Ryan Sebbane, and Morgan Sénéchal — designed a generative AI capable of summarizing information from a given document corpus. This chatbot can produce texts, tables, or images, and can be deployed in businesses or hosted on the cloud. One of its major strengths is its rapid implementation: it takes only 17 minutes to become operational.

The tool was developed using open-source language models available on the Hugging Face platform. The students are currently working on developing their own language model for better integration of their solution. The project is a candidate in the Engineering 4.0 category of the Grand Challenge of engineering schools, with winners to be announced on May 15.

An AI Tailored to Business Needs

The AI developed by Efrei students stands out for its ability to provide sourced answers, allowing easy tracing back to the original document. For example, in the banking sector, it can provide a client's address with the exact reference of the original PDF. This feature is particularly useful in sensitive sectors like healthcare, where data protection is paramount.

The chatbot has already been tested within a consulting firm, and the students are now seeking partners to trial it in various sectors. This initiative demonstrates the utility of generative AI in the professional world, offering customized and secure solutions for information management.

The project led by Efrei students illustrates the potential of generative AI in the professional world. By offering a solution that is quick to deploy, secure, and tailored to the specific needs of businesses, this initiative could mark an important step in the integration of artificial intelligence into business processes.