Artificial intelligence as a teammate: How to achieve a successful collaboration
An increasing number of work teams are using artificial intelligence (AI) for decision-making and automation purposes. What do members of such teams think of their digital “colleagues”? How does the collaboration between human beings and AI work, and how can it be optimised so that the right decisions are made and the team goals are achieved?
Project description
By means of four studies in real teams from the fields of medicine and aviation, we will investigate the cooperation between human beings and artificial intelligence. In the first study, we will investigate mental models, team processes and team performance in medical teams working intensively with AI, and will draw comparisons with conventional teams. In the second study, we will use the results of study 1 to develop a new AI for medical teams. In the studies 3a (with medical teams) and 3b (with cockpit crews), we will develop and test an intervention training unit intended to optimise teamwork with AI. From this we will derive a generally valid online tool for the optimisation of teamwork with AI, which can be used by teams from all industrial sectors.
Background
The use of AI in a team context has not yet been widely investigated. Scientific literature tells us that shared mental models (the understanding of what is done by whom, when and how) are essential for decision making, coordination and the achievement of goals in teams. However, due to the complexity and black box problems of self-learning systems, it is not yet known whether and how these mental models also apply to AI, and what this situation means for team success.
Aim
The goals of this project are the following:
- We will explore mental models, coordination and goal achievement in teams working with AI, and will compare these teams to conventionally working ones.
- We will optimise the design of an AI for use in a team context and will promote its acceptance.
- We will develop and validate an intervention training for the optimisation of teamwork with AI.
- We will develop an online tool to be used by teams in different industrial sectors.
Relevance
The project has great practical benefit, since an increasing number of teams work with AI and are affected by the problems described above. All data is collected in real teams, which ensures the transfer from theory to practice. The application-oriented recommendations in the form of specific training programmes and online tools are of direct use to teams working in the fields of medicine and aviation, or in other industrial sectors.
Original title
From Tools to Teammates: Human-AI Teaming Success Factors in High-risk Industries