Experts at McKinsey & Co. said there are six measures government agencies should implement to mitigate the risk of bias associated with the use of advanced analytics, which could help agencies transform public services, and one of those is making a senior leader responsible for risk management related to advanced analytics models.
“In federal agencies that lack robust enterprise-risk-management structures, it could sit with the chief information officer, chief data officer, or the senior-level executive responsible for governance and oversight of technology,” the experts wrote in an article posted Friday.
“It remains the responsibility of the most senior agency leaders, however, to help translate agency mission and values, such as equity and diversity, into guidance for AI risk-management leaders,” they added.
Other measures McKinsey experts suggested are establishing and communicating a clear set of analytical practices and standards; creating algorithm review panels; appointing an analytics ombudsman who could serve as a spokesperson and point of contact for external stakeholders; strategizing at the enterprise level; and building a model-risk-management infrastructure.
To create a model-governance program, there are factors agencies should consider such as creating and maintaining an inventory of models in use across the agency and developing the standard workflow for models to help ensure the adoption of best practices in bias reduction and awareness and data science.
Rahul Agarwal, a senior expert at McKinsey’s New Jersey office, co-wrote the article with senior adviser Mark Levonian, senior expert Eric Schweikert and partner Catharina Wrede Braden at the management consulting firm’s Washington, D.C. office.