Which tasks can be entrusted to artificial intelligence, and when should humans remain involved in decision-making processes? Determining how decision-making responsibilities should be distributed between humans and AI is one of the key challenges facing companies in the AI era.

Prof. Jan vom Brocke and Dr. Thomas Haskamp from the Department of Information Systems at the University of Münster, together with Dr. Ina Sebastian and Prof. Peter Weill from the MIT Center for Information Systems Research (CISR), have developed a framework to help companies address this question. The “AI Decision Matrix” enables executives to determine, based on two dimensions, the extent to which humans and autonomous AI agents should be involved in decision-making processes.

The researchers developed the matrix based on interviews with 30 executives. It is based on the understanding that decisions vary in complexity and risk: While some decisions are predictable and involve limited risk, others are characterized by ambiguity and can have far-reaching consequences. Depending on the nature of the decision, greater human involvement or a higher degree of AI autonomy may therefore be appropriate.

Learn more about the framework here.