Developed by the Luxembourg Institute of Science and Technology (LIST) and the University of Luxembourg’s Interdisciplinary Centre for Security, Reliability and Trust (SnT) and released through the Luxembourg AI Factory, the AI Assessment Sandbox Configurator enables organisations to test AI systems faster and more reliably.

Demonstrating that an AI system is trustworthy remains a costly and largely manual process. Identifying the appropriate tests, selecting the right metrics and setting up a suitable sandbox environment from scratch can require significant time and resources that many teams do not have.

The AI Assessment Sandbox Configurator is designed for organisations deploying or overseeing AI systems, including private equity and venture capital firms assessing AI tools within their own operations or across their portfolio companies. It allows users to assess AI systems without requiring specialised expertise in AI testing.

The tool enables organisations to configure tailored sandbox environments based on a shared, open catalogue of tests, controls and datasets. It brings assessment results together in a collaborative dashboard, with comparable outputs and tamper-proof logs to support auditability.

By providing a common assessment framework, the solution helps technical, legal and business teams develop a shared understanding of results and facilitates collaboration across different stakeholders.

Read more: AI Assessment Sandbox Configurator - LPEA

Explore the tool: Testing the solution | Luxembourg AI Factory - AI Factory