Artificial intelligence (AI) robots that use algorithms to complete source search tasks often encounter obstacles that prevent them from successfully completing their goals. To address this issue, researchers have proposed a human-AI collaboration that leverages the unique skills of the human brain. This collaboration allows humans to assist AI-driven robots in overcoming challenges that they cannot solve autonomously.
The researchers conducted a study to prove the feasibility of their human-AI collaboration strategy. They first identified different types of hazards that the robots could encounter and categorized them based on whether or not a human observer could help solve the problem. If the problem could be solved with human assistance, the AI robot would develop an explanation of the problem and seek help from humans through crowdsourcing. On the other hand, if the problem was unsolvable with human assistance, the search would be stopped.
The study also tested two control modes of the AI robot: Full Control and Aided Control. In Full Control mode, the human collaborator takes over the search process. In Aided Control mode, the problem-solving decision tree determines whether the collaboration between humans and AI would be beneficial.
The results of the study showed that participants in the Aided Control mode felt they had less cognitive workload and were able to effectively address the problem when they received information from the algorithm but did not take over complete control. However, non-experts had difficulty understanding the AI-driven robot’s explanations of the problem, suggesting the need for personalized interactions and plain language explanations based on the experience of the human collaborator.
Moving forward, the researchers plan to further explore personalized interactions based on the background, education level, and personality of the human participants. They aim to harness the benefits of human-AI collaboration in various application scenarios, such as natural language processing and image analysis.
This study highlights the potential of human-AI collaboration in improving the efficacy and efficiency of AI-driven robots, particularly in challenging and unfamiliar environments. By combining the strengths of humans and AI, the possibilities for successful source search tasks are greatly enhanced.
Source: Journal of Social Computing, Tsinghua University Press