Technical Committee (TC) Methods for Testing and Specification (MTS)
We create standards related to testing and specification languages and provide frameworks and methodologies to enable the other ETSI committees to achieve this goal. Work is performed very closely with ETSI’s Centre for Testing and Interoperability (CTI) to develop the background material which they then use in their support of other ETSI committees as well as other relevant standardisation bodies such as ITU Study Group 17. Much of work done by TC MTS has also been adapted and used beyond ETSI by other organisations, fora, and industry globally.
MTS AI Working Group
Artificial Intelligence (AI) is increasingly embedded in critical infrastructures, business processes, and consumer applications. Ensuring the trustworthiness of these systems through standardised methodologies for testing, documentation, and auditing is essential to building confidence, enabling interoperability, and ensuring regulatory compliance. At the same time, AI itself is becoming a powerful enabler for automation in testing and auditing, supporting activities such as automated test generation, test data creation, evaluation of execution results, and continuous monitoring.
The ETSI MTS Working Group AI ("ETSI MTS AI"), operating under TC MTS, contributes to the future of AI standardisation by developing methodologies, tools, and specifications that support trustworthy, testable, and auditable AI systems across all lifecycle stages, while also exploring the use of AI techniques to enhance the efficiency, consistency, and scalability of testing and auditing processes.
Tasks and Responsibilities
MTS AI develops methodologies and specifications to:
- Support testing, documentation, certification, and assessment of AI-enabled systems.
- Provide frameworks for conformance, interoperability, and trustworthiness evaluation.
- Enable continuous conformity assessment and lifecycle-integrated quality management.
- Collaborate across ETSI and with other SDOs to ensure consistency, interoperability, and global impact.
The technical work addresses the development of advanced methods for specification, testing, and quality assurance of AI, leveraging AI to automate testing activities such as test generation, data creation, execution optimisation, and results documentation.
Achievements and Ongoing Work
MTS AI has developed methodologies for the testing of ML-based systems, defining test types, quality criteria, and lifecycle integration for supervised, unsupervised, and reinforcement learning, as described in ETSI TR 103 910 – Test Methodology and Test Specification for ML-based Systems. Furthermore, ETSI TR 104 119 – Documentation Schemes for AI-Enabled Systems provides guidelines and recommendations for the consistent and continuous documentation of AI systems, including industrial-grade documentation techniques, use case examples, and structured documentation approach. Complementary work extends these efforts towards continuous auditing and conformity assessment (CABCA), specifying the underlying principles, roles, and procedures as set out in ETSI DTS/MTS-104008 – Continuous Auditing-Based Conformity Assessment.
Roadmap and Future Vision
In the near future, MTS AI will focus on developing industrial-grade guidelines and templates for technical documentation to ensure regulatory compliance and practical usability across sectors. At the same time, the group will advance comprehensive methodologies for AI assessment and auditing, including risk and capability evaluation as well as continuous auditing practices such as . A major part of this work is the systematic use of AI to automate conformance testing: AI techniques will be applied to generate test specifications, create and manage test data, optimise execution, and document results. This approach not only supports the assessment of AI-enabled systems and GPAI but also strengthens ETSI’s broader conformance testing ecosystem. Together, these activities will converge in a unified testing and auditing framework that equips industry with scalable, automated, and reproducible methods to demonstrate conformity through standardised deliverables, harmonised documentation schemes, and machine-readable formats.
Latest Publications
