Due to the advancement in technology, new classification models are able to sense the extent to which humans trust the intelligent machines with which they collaborate on a routine basis. This is yet another step forward towards the improvement of the quality of teamwork and human-machine interaction.
Overall, the field of artificial intelligence has had one primary goal. That is the design machines that are highly intelligent and are able to adapt to the needs of humans by changing their behavior. This will foster a greater level of trust. The new classification models in research have been developed by associate professor Tahira Reid and assistant professor Neera Jain at the school of Mechanical Engineering at Purdue University.
Intelligence systems and intelligent machines are becoming a common utility in every life of a person, according to Jain. Furthermore, trust is one of the most important factors to allow synergistic interactions between humans and machines.
For instance, industrial workers and aircraft pilots interact with automated systems routinely. In case the system starts to falter, humans override these intelligent machines without any need to do so.
According to Reid, it is a well-established fact that in order to allow successful interactions between the machines and humans, trust is the central factor.
In this regard, researchers have formed two types of empirical trust sensor models that are based on classifiers. This is a big step towards the improvement of trust between humans and machines.
This current work is fully aligned with the Giant Leaps Celebration to Pursue which acknowledges the global advancements made by the university in the field of artificial intelligence and automation, on the 150th anniversary of Purdue.
There are two techniques that are used by the model to provide data in order to gauge the trust. The first is the electroencephalography and the other is galvanic skin response.
Among these, the first technique makes use of patterns in the brainwaves and records them. At the same time, the second technique analyzes the variations in the skin’s electrical characteristics.
The findings of this research have been detailed in a special issue of Transactions on Interactive Intelligence Systems. The special issue of the journal is titled as “Trust and Influence in Intelligent Human-Machine Interaction.”
The author of this paper is Kumar Akash, a mechanical engineering graduate student along with Wab Lin Hu, who is a former graduate student now involved in postdoctoral research at the Stanford University.
According to Jain, the primary interest in this research is to design machines by using feedback-control principles. These machines should have the capability to respond to the variation in the level of trust that a human has in real time. In order to make this happen, a sensor that estimates the level of human trust in real time is required.
Furthermore, Jain added that the future is revolving around human-agent collectives that require a lot of collaboration, and efficient coordination between the machines and the humans. For instance, consider a number of robots that are helping a human rescue team in the aftermaths of a disaster. In this scenario, a number of humans and machines will have to be considered. Considering this fact, Jain hopes to scale up the work to include multiple machines and humans working together.