GAO Issues Framework to Keep Federal AI Systems Accountable

The principles were designed for the U.S. government, but in the absence of broad-based societal standards, some say they could provide a template for the private sector

The U.S. Government Accountability Office recently released its artificial-intelligence accountability framework, intended to instruct federal agencies and the vendors working with them on key principles for creating responsible and reliable AI systems.

“It’s more about what questions need to be asked to assure that an AI system is implemented in a way that’s going to achieve its goals, but also do that efficiently, effectively, ethically and equitably,” said Stephen Sanford, managing director and head of the GAO Center for...