AI technologies are typically used across an organization rather than in one department. To effectively implement Equitable AI in your organization, you will likely need to work across departments. Internally, this usually means involving people from IT, HR, Product, Compliance, Legal and other groups like those responsible for Diversity, Equity, Inclusion, and Accessibility (DEIA). In addition, you should also consider the external stakeholders who may use the AI technologies your organization implements, such as consultants, auditors, partners, customers, and prospective employees.
For your Equitable AI initiative to be successful, your work cannot be done in a vacuum. Identifying your stakeholders helps you determine who must be engaged in the Equitable AI initiative, when they must be engaged, and how engaged they should be.
As part of this process, categorizing your stakeholders into roles can help you better understand who will advise, champion, implement, and be impacted by your Equitable AI initiative. This will help make sure you build out an effective internal community and develop training resources to support your Equitable AI initiative.
Tips to Get Started
- Use the tipsheet Staffing for Equitable AI: Roles and Responsibilities to consider the different stakeholder roles you will engage in your organization. Different stakeholder categories include:
- Advisors are senior executives, business managers, and technology leaders—including any internal subject matter experts, and external consultants or researchers—who will build, lead, and advise your organization’s Equitable AI initiative. Examples: Chief AI Ethics Officer, CEO, Board of Directors.
- Champions are the senior leaders and influencers who will promote the Equitable AI initiative across your organization. Examples: Employee Resource Groups (ERGs), Diversity, Equity, Inclusion and Accessibility (DEIA) Initiatives.
- Implementers are the people who will be responsible for executing your Equitable AI processes on a day-to-day basis. This includes people who will plan, configure, integrate, and monitor AI technologies your organization deploys. Examples: internal employees, contractors, external consultants.
- Impacted End Users are the people who can ultimately benefit from having Equitable AI tools in place—whether by using them directly (e.g., communicating with a chatbot) or indirectly (e.g., receiving highlights from automated data analytics). End users might include employees performing their jobs, as well as prospective employees, customers or business partners who are accessing your organization’s AI-enabled technologies.
- Use PEAT’s Equitable AI Stakeholder Mapping Template to outline teams and individuals who fit into stakeholder categories like Advisors, Champions, Implementers, and Impacted End Users.
- Make sure to include both internal and external teams and individuals you think should be engaged in your initiative. You’ll also need to determine the best ways to reach each type of stakeholder.
Tips for Ongoing Operations
- Maintain your stakeholder map. It’s important to keep your stakeholder mapping current by continuing to document your Advisors, Champions, Implementers, and Impacted End Users. Make sure that, as your Stakeholder Map evolves, you communicate changes and make it available to your entire organization.
- Are there existing senior leaders or managers who support or lead AI efforts in your organization and who have an interest in equity and fairness in AI?
- Who are the key groups and leaders you think might champion your initiative—the experts who will lend their credibility and influence to help embed Equitable AI in your organization?
- What internal roles exist to help plan, execute and monitor AI implementations? Are contractors and external consultants involved?
- How will you describe the value of participating in your initiative in your invite and when you eventually meet with stakeholders?
- How will you bring your community together? Do some participants already know each other?
- Which accessible communication channels will you use for collaboration?