The NVIDIA Halos AI Systems Inspection Lab is ANAB’s first accredited inspection program for the safety of physical AI systems. Accredited to ISO/IEC 17020, the international standard for inspection bodies, the lab provides companies building with NVIDIA Halos a structured means of evaluating their systems against recognized safety requirements as a foundation for future certification activities.
The accreditation marks an important step in developing trusted assurance mechanisms for AI-enabled technologies operating in the physical world. As robots, autonomous systems, and intelligent machines move into industrial, commercial, and shared public environments, stakeholders need confidence that safety claims are evaluated through competent and impartial processes.
NVIDIA Halos for Robotics combines safety-focused hardware, software, and inspection capabilities designed to support the development and deployment of physical AI systems. ANAB's accreditation provides independent verification that the inspection program operates in accordance with internationally accepted requirements for competence, consistency, and impartiality.
Unlike generative AI systems that operate primarily in digital environments, physical AI systems interact directly with people and infrastructure, creating unique safety challenges. As regulatory frameworks and industry expectations continue to evolve, independent conformity assessment will play an increasingly important role in building confidence in these technologies.
“AI safety evaluation only carries weight if the bodies performing it can demonstrate competence and impartiality against recognized standards,” said Doug Leonard, executive director of ANAB. “Accreditation provides that demonstration. Extending it to physical AI inspection gives manufacturers, regulators, and the public a credible, common basis for trusting safety claims as this technology moves into wider use.”
This accreditation reflects ANAB's broader commitment to supporting trustworthy AI through internationally recognized conformity assessment programs. As AI governance and safety frameworks continue to develop, accreditation offers an established way to demonstrate confidence in the organizations responsible for evaluating these systems.