����ֱ��

New Research Identifies Ideal AI System Types for Managers

As managers face more pressure in implementing artificial intelligence (AI) into the workflow, a study from researchers at ����ֱ�� and two other schools offers insights to help managers adapt.

---

As managers face more pressure in implementing artificial intelligence (AI) into the workflow, a study from researchers at ����ֱ�� and two other schools offers insights to help managers adapt.

By distinguishing between managers who are risk-averse and those that are more risk-taking, the study sheds keen insight into how managers might delegate tasks to AI, the type of AI they are more likely to adopt and risks inherent in those systems.

“As a manager, you need to be willing to give away some autonomy as the AI systems are making some of these decisions in your workflow process,” said , Ph.D., associate professor of information systems in FAU’s . “Some managers are more willing to delegate, and others are less so. Depending on their manager profile, it will affect how they invest in AI systems that can act and learn autonomously. What is at stake is the generation of business value when decisions are delegated to an AI.”

The conceptual paper “” was published in the Journal of Management Information Systems with Abhijith Anand, Ph.D., assistant professor at the , and Aaron Baird, Ph.D., associate professor in the Institute for Insight at the

The paper breaks down managers into three categories based on their management styles for investment decisions: projective, iterative, and practical evaluative. Managers with a projective temporal tone are more future-oriented and focused on experimentation and innovation. Managers with an iterative temporal tone are more past-oriented and concerned with established protocols and routines, while those with a practical evaluative temporal tone fall somewhere in the middle - the present demands and needs of the business are at the forefront of their minds.

More future-focused managers might be better off investing in AI that push the boundaries on autonomous learning and are able to adapt autonomously based on incoming data. This can range from augmentation, where the AI system learns from and complements a human action for a given task, to automation, where the AI is a full substitute for human actions on a given task.

For risk-averse managers, an AI system that learns autonomously and make decisions on its own may seem too risky. These managers are likely to find predictable AI systems to be more beneficial as they maintain some control over AI-based decision-making.

“The investment choices depend on the manager’s delegation preference and how much control they are willing to release, what their value creation goals are to decide on the type of AI they want to invest in,” Queiroz said. “A more advanced AI system might be riskier and scarier for some managers since they have to give away some agency to a system that could replace them in the future.”

-FAU-

Tags: