Strengthening confidence in its AI solutions for Sopra Steria
Ingrid FIQUET, Manager Data & IA, Sopra Steria Next
“The Confiance.ai program has provided us with a complete chain for verification and transition to industrial production, combining methodology and technical tools. Proven on the program’s wide variety of use cases, it now enables us to address most of our customers’ needs.”
By contributing to the Confiance.ai program, Sopra Steria has strengthened the reliability and effectiveness of its AI solutions by integrating trusted components and developing robust methods. These advances support the company’s competitiveness and innovation, while ensuring greater compliance with future regulations.
AI adoption issues specific to Sopra Steria
Due to the diversity of its customers, Sopra Steria is confronted with very different fields of AI application, each with its own constraints. It is essential to have the methodological elements and technical tools to determine and verify the trust parameters adapted to these varied situations.
Products and services affected by the Confiance.ai program
Sopra Steria’s Data & AI platform has been impacted by the integration of trust components and the evolution of the MLops technical environment. These components have also influenced customer support, bringing a trusted AI dimension to industrialization missions for AI projects.
Confiance.ai’s contribution to Sopra Steria’s strategy and challenges
Confiance.ai has contributed to the industrialization of AI, a subject at the heart of Sopra Steria’s strategy. The program has made it possible to build MLops environments and integrate trust components into AI use cases.
Competitive advantages developed with Confiance.ai
Sopra Steria has strengthened its convictions and service offering on the construction of industrial AI, by integrating trust bricks to prepare for the implementation of the AI Act. Trust components have been integrated into an offer proposed on the AWS Marketplace, based on the generation of synthetic images from LLM for the detection of objects in critical scenes.
Challenges and conditions for success for the next phase of the Confiance.ai program
It is crucial to continue the program to increase the maturity of the components in the Confidence catalog, in order to increase the number of components that can be used in a real industrial context. It is also important that Trust results are further marked to comply with the AI Act.
Example of current or future implementation/deployment
Sopra Steria has developed an end-to-end demonstrator based on the Renault use case, using a dataset of tires to illustrate the transfer of trust components, notably in terms of robustness, inferential image validity and explicability.
Accelerating AI Act compliance with the Confiance.ai program
Confiance.ai provided technical solutions, in-depth knowledge of trusted AI concepts and methodological leads, enabling us to meet the requirements of the AI Act.
