Norges Bank Investment Management (NBIM) expects the use of AI to become commonplace, if not necessary, and expects its portfolio companies to use AI responsibly. From NBIM’s perspective, responsible use encompasses board accountability, explainability and transparency, and robust risk management.
Board accountability—Board oversight is effected by directors having the requisite expertise and resources to oversee this area as relates to the company’s risk exposure and business model specifically, and includes: (i) oversight of business-specific AI policies, guidance, and governance structures that are aligned with internationally acknowledged standards for human rights and responsible business conduct, relevant AI guidelines, and sector-specific tools and best practices; and (ii) oversight of the corporate culture surrounding the implementation of responsible AI.
Explainability—Companies should be able to explain how their AI systems have been designed, trained, and tested and how they align with human values and intent, and should provide information to “relevant” third parties such as auditors or regulators to allow for third-party verification and risk assessment.
Transparency—The company should be transparent about its use of AI with stakeholders so that they know whether, when, and how it is being used.
Robust risk management—Risk management processes should be consistent with the company’s risk exposure and should address a wide range of potential risks (including privacy, security, misinformation, deception, discrimination, and biases) relative to the company specifically, as well as to stakeholders and the society at large.
See “Norway wealth fund to firms: use AI, but do it responsibly” (Reuters) and additional resources on our Artificial Intelligence page.
This post first appeared in the weekly Society Alert!