DEC 19, 2018

Classification Model Measures Human Trust in Intelligent Machnines

WRITTEN BY: Nouran Amin

Research at Purdue University proposes new "classification models" capable of sensing how well humans trust the intelligent machines they interact with. "It is well established that human trust is central to successful interactions between humans and machines," says associate professor of mechanical engineering at Purdue University, Tahira Reid. The goal of the research is to enhance intelligent machines with new design capability for improving how these machines adapt to human trust by changing their behavior.

How should intelligent machines be designed so as to "earn" the trust of humans? New models are informing these designs.

Credit: Purdue University photo/Marshall Farthing

"Intelligent machines, and more broadly, intelligent systems are becoming increasingly common in the everyday lives of humans," says assistant professor Neera Jain in the Purdue University School of Mechanical Engineering. "As humans are increasingly required to interact with intelligent systems, trust becomes an important factor for synergistic interactions."

The classification model uses two techniques, electroencephalography and galvanic skin response, that provide data to gauge trust. Electroencephalography records brainwave patterns while the galvanic skin response monitors the change of electrical characteristics in the skin--providing the psychophysiological "feature sets" that correlates with trust.

Results of the study using human subjects were outlined in the Association for Computing Machinery's Transactions on Interactive Intelligent Systems. "We are interested in using feedback-control principles to design machines that are capable of responding to changes in human trust level in real time to build and manage trust in the human-machine relationship," says Jain. "In order to do this, we require a sensor for estimating human trust level, again in real-time. The results presented in this paper show that psychophysiological measurements could be used to do this."

Learn More:

.

Findings concluded that the classification model effectively induced trust and distrust in intelligent machines. "In order to estimate trust in real time, we require the ability to continuously extract and evaluate key psychophysiological measurements," Jain said. "This work represents the first use of real-time psychophysiological measurements for the development of a human trust sensor."

Source: Purdue University