Our analysis of board responsibilities for artificial intelligence oversight is based on a global standard for corporate governance practices — the G20/OECD Principles of Corporate Governance. The Principles, first released in 1999 and updated in 2004 and 2015, is a set of guidelines for boards of directors adopted by the OECD Council and endorsed by the G20 Leaders. The 2015 version of the Principles was developed and reviewed by the OECD Corporate Governance committee, with input from experts from the Basel Committee on Banking Supervision, the Financial Stability Board and the World Bank Group.
To prepare the AI oversight toolkit, the Forum’s Centre for the Fourth Industrial Revolution in San Francisco analyzed the board responsibilities described in the Principles and applied them to artificial intelligence. The analysis was reviewed by legal experts in corporate governance at Centre member companies and tested by directors at boards that piloted the toolkit.
The Principles suggest responsibilities that are equally applicable to many of the 12 modules in the AI oversight toolkit. To avoid repetition, we have gathered these common responsibilities here in the Responsibility module. Principles and AI responsibilities that pertain only to one module are described in the respective module.
Values and ethics should be integrated into all business decisions, rather than be treated as an afterthought. We agree with the authors of the G20/OECD Principles that:
High ethical standards are in the long-term interests of the company as a means to make it credible and trustworthy, not only in day-to-day operations but also with respect to longer-term commitments. (Principle VI:C)
Accordingly, in this module and others, we include ethical responsibilities with others rather than treat them separately. The Ethics module focuses on setting of ethics standards.
The analysis in this module does not constitute legal advice and is not intended to address the specific legal requirements of any jurisdiction or regulatory regime. Boards are encouraged to consult with their legal advisors in determining how best to apply the principles discussed in this module to their company.
Access to information
In order to fulfil their responsibilities, board members should have access to accurate, relevant and timely information. (Principle VI.F)
To ensure access to information required to oversee AI’s application to strategy:
- Board members should be able to obtain the information needed to fulfil their oversight responsibilities relating to AI from management, other employees and third parties,
- Board members should have access, at no cost to them, to timely advice from qualified advisors.
Performance, expenditures, acquisitions
The board should fulfil certain key functions, including… setting performance objectives; monitoring implementation and corporate performance; and overseeing major capital expenditures, acquisitions and divestitures. (Principle VI.D.1)
AI-related topics that boards should review and guide may include:
Setting performance objectives, monitoring implementation and performance:
- The alignment of the company’s AI use with its strategy, its key performance indicators, and its key risk indicators.
- The impact of AI on management’s shareholder value indicators and other shareholder objectives.
- Progress towards successful implementation of AI plans.
- Whether and how to factor AI performance into performance objectives for management.
- Adherence of suppliers and partners to the company’s AI ethics principles.
- The effectiveness of management’s use of AI to accelerate key processes and improve productivity.
Major capital expenditures, acquisitions and divestitures:
- Major investments in AI systems, services and talent.
- The impact of company’s acquisitions on its ability to use AI.
- New business and ethical risks introduced by acquired companies.
- Assurance that acquired organizations follow the company’s AI ethical standards.