Shaping the future of AI assurance: DNV Ventures brought the ecosystem together in San Francisco

While AI is accelerating into every sector of the global economy, 95% of industrial AI pilots fail because they can`t be trusted to scale safely. This demonstrates that trust in AI is no longer optional; it's essential, forming the foundation for deploying AI into high-risk assets and critical applications.

Earlier this month in San Francisco, DNV brought together leading voices in AI - including global researchers, AI investors, entrepreneurs, and technical experts - for the launch of DNV’s 2026 Position Paper, “Assurance of AI‑Enabled Systems.” The event sparked intense and constructive dialogue on shaping the future AI assurance industry, with participation from key players such as OpenAI, AI Trust Frontier Labs, and scientists from UC Berkeley. 

A key takeaway from this paper is that the most important step towards trustworthy AI is to consider it as part of a system. For example, a robotaxi is more than a car; it is an AI-enabled system. When AI fails, it is most often more than a ‘glitch’, but the result of hidden risks and biases, unclear responsibilities, or unexpected interactions between AI, people, and the environment where it operates. Fortunately, proven methods exist to assure AI-enabled systems that point the way towards a new paradigm for assurance as a continuous, adaptive, system-wide, and evidence-based process, beyond static checks.  

The system view of AI also underlines that AI safety cuts across sectors and demands collaboration across industries, disciplines, and between academia, industry and regulators.   

To support this shift, DNV is deploying capital, scientific research, and global partnerships to help shape the AI Assurance Industry. DNVs investment arm, DNV Ventures, is among the first global investors actively investing in startups focused on AI trust.  

DNV is also expanding global partnerships, including the University of York and the Centre for Assuring Autonomy. We work with NASA on uncertainty quantification. We also have a valued partnership with the Center for Information Technology Research in the Interest of Society (CITRIS) at UC Berkeley on R&D in embodied AI. 

This commitment is a natural extension of DNV’s legacy of building trust through every major industrial and technological transition. DNV believes AI can only be adopted and scaled safely when it's trusted, and we are committed to preventing systemic failures and protecting society and industry as AI becomes deeply embedded into critical systems.