Quality Assurance for AI
AI technology adoption is experiencing rapid growth, having tripled between 2017 and 2021 according to McKinsey. Alongside this growth, the risks associated with AI incidents, specifically ethical biases, prediction errors, and cybersecurity threats, are on the rise. However, the existing AI quality tools are inadequate and rely heavily on manual testing, creating a gap that AI/ML engineers cannot fill in terms of workload, costs, and demand within the required timeframe.
To address this pressing need, GISKARD is developing an open-source and SaaS (Software-as-a-Service) solution tailored for companies seeking quality assurance for their AI models. GISKARD’s software platform enables automated AI Quality Testing, Inspection, and Remediation. As an AFNOR member, which is the French national standards council, GISKARD is dedicated to becoming the leading European software provider, helping organizations prepare for the forthcoming EU AI Act.
The support of the EIC (European Innovation Council) is essential to achieve this goal. In this project, GISKARD aims to optimize and validate its AI Testing solution while expanding its applicability to additional use cases such as Time Series and Computer Vision.