IMT, France’s leading public group of engineering and management Grandes Écoles, is taking a new step forward in its schools’ AI experimentation programme.
- It provides Le Chat licences to train and equip its staff, as well as extended API access for its technical experts.
- Under the terms of their agreement, IMT and Mistral AI will also organise a hackathon for students from all of the group’s schools.
- IMT has thus become the public higher education institution that has invested the most in the technologies developed by Mistral AI.
- The IMT is also joining the ILaaS federation to pool IT resources and secure data usage alongside the twenty or so higher education institutions already involved.
- Finally, through its AI programme, the IMT aims to set an example and ultimately ‘certify’ the ethics of its practices in terms of data and sober artificial intelligence. This is the purpose of an announced collaboration with Ekitia to design experiments that meet the requirements of its label.
- All these partnership initiatives aim to equip IMT schools to experiment, teach and innovate with sovereign and high-performance generative AI models.
Mistral AI : IMT accelerates its efforts in sovereign AI
In the context of the development of generative AI, the IMT is taking a major step forward with the deployment of Le Chat Mistral AI licences (French open source models) that will enable the collective of schools to train internally, create or test new use cases across all activities: teaching, research, support and assistance functions. The agreement with Mistral also provides access to extended APIs for hundreds of researchers and expert users, such as information systems departments.
Attentive to the profiles of students trained by the group’s schools, Mistral AI is also involved in the joint organisation of a national inter-school hackathon, following on from the many IMT hackathon experiences shared with large companies.
ILaaS : a new alliance for shared usage
The IMT has also formalised its entry into the Ilaas (Inference LLM as a Service) mutualised federation, which brings together twenty higher education institutions to develop trustworthy, robust and sober generative AI by pooling technical resources. The members of the federation pool IT resources and also experiment with use cases common to ESR (higher education and research) in a collaborative approach. This initiative enables IMT to access and contribute to a shared infrastructure, participate in its governance and test new uses in a common collaborative framework.
Ekitia label : trust and data orchestration for experiments
In order to eventually ‘certify’ the ethics of data use in its AI applications, the IMT aims to roll out several innovative experiments designed to meet the requirements of the Ekitia label. This label and the associated charter aim to ensure ethical and sober use of data. The orchestration and sharing of data from multiple environments (school information systems, business software, open warehouses, etc.) will thus be studied collaboratively between Ekitia and IMT stakeholders in order to deploy new use cases with guarantees of transparency, trust and responsibility for users and stakeholders.
Cécile Dubarry, Director General of IMT : “We want to make artificial intelligence a lever for 360° innovation. By focusing on Mistral AI tools, pooling resources via ILaaS and the Etikia label, we are strengthening our collective capacity to respond to educational, scientific and digital challenges. “
Lucie Jacquet-Malo, AI programme coordinator at IMT ; “These structuring partnerships enable IMT’s schools to fully play their role as a living laboratory for responsible AI. They strengthen our ability to structure a national and European dynamic serving multiple
