Accountability is the state of, or duty on, a person or organization to answer for actions, decisions and results, including decisions an AI system makes, and using AI can change existing accountability arrangements (ISO/IEC 23894 A.2 and ISO/IEC 38507 give more). For each identified interested party the organization should analyse the foreseeable benefits and harms tied to accountability, such as: effects on the party's legal status, life opportunities or human rights, and the possibility of physical or psychological harm; effects when the system does not perform as intended; effects arising from how data used to develop the system was chosen and used; effects of missing or inadequate human oversight of the system and its outputs; and effects of being unable to reverse or challenge the system's decisions.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.