Skills for Collaborative Intelligence
Understanding the role and experience of the human in optimising collaborative intelligence.
The Challenge
Collaborative intelligence utilises the combined strengths of human and artificial intelligence. Much effort has been directed towards the technical challenge involved in creating an AI agent that is capable of collaborating with a human on a shared task. But an effective collaboration depends on both collaborators. In this project we investigate what human skills, knowledge, aptitudes and work design choices promote effective human AI collaboration.
Our Response
AI can reduce some of the variability in human performance by providing specific knowledge and capability that is needed for a task. However, collaborative AI will make other human skills, knowledge and aptitudes more importance because it depends on two-way interaction between the human and the AI on a shared task. The ability to direct and give feedback to the AI and respond to the output of the AI is likely to be critical for effective human AI collaboration. In addition, the improved situational awareness and conversational capability of the AI provides more choices for the human worker in terms of how and when the AI is deployed, increasing the need for metacognitive skills and critical thinking. Working with collaborative AI may also affect task performance by altering motivational work characteristics such as autonomy, variety and social connectedness so this research also explores the human experience of collaborating with AI.
It is important to acknowledge that the skills needed to work with collaborative AI may vary according to the task being performed in the collaboration, the environment in which the collaboration occurs (physical or virtual for example) and the way in which the AI is designed and implemented. In response to this complexity, we adopt a variety of approaches:
- Interviewing workers and managers to understand what characteristics differentiate (a) workers who use collaborative AI tools more effectively and (b) effective and ineffective interactions with collaborative AI
- Testing the effect of potentially relevant skillsets (e.g., metacognition, AI literacy and critical thinking) in a range of contexts
- Examining how work design choices (associated with the implementation of collaborative AI) affect motivational characteristics and ultimately, collaborative performance
Impact
This project focuses on the role of humans in optimising collaborative intelligence, counterbalancing research which focuses on technological capability. By decreasing the importance of some skills and knowledge and increasing the importance of others, collaborative AI has the potential to change the source of competitive advantage for human workers. This research supports the AI-readiness of the Australian workforce by providing early insight into the human factors that differentiate effective human-AI collaboration.
News
Newsletter
If you’d like to stay updated on our research through monthly newsletters, click the button below to subscribe.
External Collaborators
Queensland University of Technology and Australian Cobotics Centre
Curtin University
Aged Care Industry Information Technology Council
Students
Yijing Liao
- Understanding how Generative AI influence task performance: the role of skill level and human learning
Sichen Meng
- Enhancing Job Satisfaction and Efficiency in Warehouses through Cobots Integration
Phuong Anh Tran
- The application of collaborative robots in manufacturing: A study of work design and motivational potential
Other Initiatives
The skills team participated in the second Science Digital/Sigma8 sprint, helping to develop the survey exploring use of Generative AI tools. The team provided measures of individual characteristics that have the potential to influence (a) the way in which such tools are used and (b) the benefits experienced from using these tools.