Create an AI council
It’s essential for organizations to establish a governance framework that ensures the responsible and ethical use of AI across their operations and stakeholders.
One way to achieve this is to create an AI Council, a cross-functional and multidisciplinary body that oversees and guides the development, deployment, and evaluation of AI solutions within the organization. An AI Council can help to align the organization's AI strategy with its vision, values, and goals, as well as to identify and mitigate potential risks and harms of AI. An AI Council can also foster a culture of trust, collaboration, and innovation among the AI practitioners, users, and beneficiaries within and outside the organization.
The Objectives and Functions of an AI Council
The main objective of an AI Council is to ensure that the organization's AI solutions are aligned with its ethical principles, legal obligations, and social responsibilities. To achieve this, an AI Council performs the following functions:
- Define and communicate the organization's AI vision, values, and policies.
- Review and approve the AI use cases and projects proposed by the different business units and teams within the organization.
- Model trusted leadership behavior for driving change.
- Monitor and evaluate the performance and impact of the AI solutions deployed by the organization.
- Provide guidance and support to the AI practitioners and users within the organization.
- Engage and collaborate with the external stakeholders of the organization.
- Promote and facilitate the learning and innovation of AI within the organization.
The Structure of an AI Council
An AI Council should include the following elements:
- IT enablement team. Responsible for technical preparedness and rollout, providing onboarding support, as well as managing feedback and compliance.
- Executive sponsor. Drives adoption and infuses confidence in the technology. Active champion that encourages integrating AI into day-to-day processes.
- Change management team. Bridge between the AI council and employees, helping to monitor adoption progress, gather feedback, and foster a collaborative environment.
- Risk management. Ensures compliance with AI regulations and ethical standards. Ensure organization’s AI initiatives are transparent, accountable, and trustworthy.
Tip
Use our AI Council Guide to learn more about the roles and responsibilities of an AI Council and help you create your own.
Responsible AI Standard
At Microsoft, we created our own framework, the Microsoft Responsible AI Standard, to guide how we build AI systems responsibly and ethically.
Creating a Responsible AI Standard helps in the following ways:
Records our practice of Responsible AI by Design – the proactive ways in which we guide the design, build, and testing of AI systems.
Establishes a durable framework for the maturing practice of responsible AI and evolving regulatory requirements.
Reflects our deeper exploration of what our six AI principles mean and the steps we must take to uphold them.
- Fairness
- Reliability & Safety
- Privacy & Security
- Inclusiveness
- Transparency
- Accountability
The Responsible AI Standard sets out our best thinking on how we'll build AI systems to uphold our core principles and earn society’s trust. It goes beyond broad concepts, offering our teams detailed, actionable guidance to implement these principles effectively in our AI initiatives.
The Standard details concrete goals or outcomes that teams developing AI systems must strive to secure. These goals help breakdown a broad principle like ‘accountability’ into its key enablers, such as impact assessments, data governance, and human oversight. Each goal is then composed of a set of requirements, which are steps that teams must take to ensure that AI systems meet the goals throughout the system lifecycle. Finally, the Standard maps available tools and practices to specific requirements so that Microsoft’s teams implementing it have resources to help them succeed.
The following image describes Microsoft’s Responsible AI Standard’s goals.