Envisioning the future of AI
AI technologies are improving and innovating. New algorithms and techniques will require less data and computing power; others will combat online trolling and improve security, create chatbots that converse better with people, find more relevant documents when it searches for them, and create AI systems that make decisions that are simultaneously more profitable and fairer. These powerful AI algorithms will be able to run faster, without requiring clouds and telecom connections, due to new hardware that is optimized for AI. AI innovation at lower cost opens new opportunities to serve customers, improve operations and increase profitability. Leaders need to stay informed of AI’s new capabilities in order to foresee AI’s future impact.
Major purchases and partnerships
While some AI software and datasets are inexpensive or even free, enterprise-level AI requires major investment in software, services, infrastructure, tools and talent. These AI systems ought to work well with a company’s existing systems and data. But if its current technology constrains a company’s use of AI, leadership need a viable plan for modernizing the infrastructure. Inevitably, companies deploying AI will forge partnerships with AI providers and consultancies. Leaders should be confident that their vendors will deliver on time and can keep promises. Additionally, they need to consider whether sharing data and knowledge with an AI provider opens the door to a future competitor.
AI use by IT
AI is already helping IT and engineering organizations to run better. The most frequent use of AI by companies has been to improve the performance of the IT and engineering functions: automating tasks such as cleansing data, and for quality control and cybersecurity. AI can also improve systems development and user interfaces. Since all organizations depend on IT, it’s important for IT departments to take advantage of AI to improve their productivity and the quality of their work, find the talent they need to employ AI and fulfil the responsibilities that come with using and building AI systems.
While the G20/OECD Principles of Corporate Governance do not specify technology within the listed responsibilities, boards cannot carry out their oversight duties without considering how their companies use and manage technology and what are their management’s major technology plans, investments and partnerships.
Many responsibilities that apply to other modules are relevant to technology:
To act in good faith, with due diligence and care, boards should be fully informed about plans for applying AI in their strategy, AI’s alignment with core values and ethical standards, the risks associated with their AI strategy, and the regulations affecting the use of AI. Directors should have access to accurate, relevant and timely information.
To oversee corporate strategy, major plans of actions, risk management and budgets and business plans, boards should review and guide management’s vision, goals, actions and expenditures for AI, its support for innovation and using new AI resources, its awareness and plans for legal compliance and ameliorating AI risk, and competitors’ use and plans for AI.
To oversee corporate performance, expenditures and acquisitions, boards should review and guide AI’s alignment with strategy, shareholder values, ethics, performance and risk indicators, implementation of AI plans, the effectiveness of AI to accelerate processes and improve productivity, major investments in AI systems and talent, and acquisitions.
See the Responsibility module for more details about the G20/OECD Principles and the above responsibilities.
To carry out these responsibilities, boards – in particular, technology committees – should also review and guide these technology-specific concerns:
Act in good faith, with due diligence and care
- Develop realistic expectations about AI, its capabilities and its risks
- Be prepared to question and, if necessary, challenge technology executives and providers about the likelihood and technical requirements for AI success
- Understand the causes of AI project failures and overruns.
Oversee corporate strategy, major plans of actions, risk management and budgets and business plans
- Identify new benefits as AI technology as its capabilities evolve
- Establish governance guidelines for AI technology, and the acquisition, use and ownership of data used in AI systems
- Guard against the wrongful use of AI technology by ill-intentioned actors within and outside the company
- Implement appropriate measures to manage human safety, data privacy and cybersecurity risks at acceptable levels
- Understand and adequately address AI-specific risks, including the risks of inaccurate AI predictions, flawed decisions made by AI systems and robots, and misguided human interactions with AI systems
- Understand and adequately address the root causes of AI-specific risks, including lack of transparency and accountability, and biased, inaccurate or otherwise flawed data.
Oversee corporate performance, expenditures and acquisitions
- Are AI services and systems purchase decisions carried out fairly and competitively, without bribes, kickbacks or other corrupt practices and influence?
- Is AI the right solution/product for the problem?
- Do AI investments and activities follow priorities set by management?
- Does the IT organization have the resources needed to manage, maintain and upgrade its AI systems and the infrastructure on which it rests?
- When is an AI-based product feature safe enough to go public?
- Are users and providers of data used in AI systems managing data responsibly?
The following suggestions can help the individual who prepares the board discussion and sets the agenda on managing AI investments and exploring the future potential of this new technology.
Before leading the first meeting
- Prepare yourself: Become familiar with AI, what it can do today and what it will be able to do in the future as the field advances. Separate the hype from reality by looking at the research and the sources behind the claims, and the issues that complicate the implementation of the technology. The Resources section provides further reading about AI technology. Speak to senior IT, security and public affairs executives about the ethics issues that are on their minds.
- Gauge board-member interest in AI technology issues: Speak to other board members. Learn what importance they place on the future of AI and the concerns they have about planned AI investments and partnerships. Identify the board members who are most interested in moving forward with new AI investments, and those who have concerns or lack interest.
- Set goals: Think ahead about the desired outcomes of the board discussion.
Set the initial agenda items
To learn about the current state and future of AI technology, steps include:
- Present: Arrange a briefing on AI. What is it? How is it already being used by your company or by companies in your business environment and industries wrestling with similar business issues to yours? What are the new capabilities of AI that are expected to emerge in the next five years? The briefing should also cover AI implementation issues and responsibilities, including ethics, and the current AI activity inside your own company.
- Discuss: Next steps for the company, whether to study the benefits and risks of technology, identify initial use cases, launch pilots or expand AI’s use inside the company.
- Delegate: Decide which members of the executive team will be responsible for carrying out the next steps, what the board expects them to achieve and when they will report back.
- Engage: Decide how the board will stay up-to-date with developments in the field of AI.
Set follow-up agenda items
These can include:
- Technological innovation: Discuss how the company can encourage AI innovation inside the IT organization, and through collaborative efforts with employees, customers, technology providers, research organizations and other outside institutions.
- Build, buy or outsource: Review management’s AI procurement strategy and whether it delivers the applications, results, flexibility and cost structure the company requires.
- Vendor partnerships: Review major AI providers, and investigate whether they are delivering on their promises or posing competitive risks.
- Emerging vendors: As part of a conversation on acquisition targets, consider which technology companies are suitable targets for acquisition or investment.
- Technology risks: Discuss major AI technical and implementation risks that could lead to significant corporate risk vulnerabilities.