What does it take for a human to trust a robot? That is what Army researchers are uncovering in a new study into how humans and robots work together.
Research into human-agent teaming, or HAT, has examined how the transparency of agents–such as robots, unmanned vehicles or software agents–influences human trust, task performance, workload and perceptions of the agent. Agent transparency refers to its ability to convey to humans its intent, reasoning process and future plans.
New Army-led research finds that human confidence in robots decreases after the robot makes a mistake, even when it is transparent with its reasoning process. The paper, “Agent Transparency and Reliability in HumanRobot Interaction: The Influence on User Confidence and Perceived Reliability,” has been published in the August issue of IEEE-Transactions on Human-Machine Systems.
To date, research has largely focused on HAT with perfectly reliable intelligent agents–meaning the agents do not make mistakes–but this is one of the few studies that has explored how agent transparency interacts with agent reliability. In this latest study, humans witnessed a robot making a mistake, and researchers focused on whether the humans perceived the robot to be less reliable, even when the human was provided insight into the robot’s reasoning process.
“Understanding how the robot’s behavior influences their human teammates is crucial to the development of effective human-robot teams, as well as the design of interfaces and communication methods between team members,” said Dr. Julia Wright, principal investigator for this project and researcher at U.S. Army Combat Capabilities Development Command’s Army Research Laboratory, also known as ARL. “This research contributes to the Army’s Multi-Domain Operations efforts to ensure overmatch in artificial intelligence-enabled capabilities. But it is also interdisciplinary, as its findings will inform the work of psychologists, roboticists, engineers, and system designers who are working toward facilitating better understanding between humans and autonomous agents in the effort to make autonomous teammates rather than simply tools.
This research was a joint effort between ARL and the University of Central Florida Institute for Simulations and Training, and is the third and final study in the Autonomous Squad Member project, sponsored by the Office of Secretary of Defense’s Autonomy Research Pilot Initiative. The ASM is a small ground robot that interacts with and communicates with an infantry squad.
Prior ASM studies investigated how a robot would communicate with a human teammate. Using the Situation awareness-based Agent Transparency model as a guide, various visualization methods to convey the agent’s goals, intents, reasoning, constraints, and projected outcomes were explored and tested. An at-a-glance iconographic module was developed based on these early study findings, and then was used in subsequent studies to explore the efficacy of agent transparency in HAT.
Researchers conducted this study in a simulated environment, in which participants observed a human-agent Soldier team, which included the ASM, traversing a training course. The participants’ task was to monitor the team and evaluate the robot. The Soldier-robot team encountered various events along the course and responded accordingly. While the Soldiers always responded correctly to the event, occasionally the robot misunderstood the situation, leading to incorrect actions. The amount of information the robot shared varied between trials. While the robot always explained its actions, the reasons behind its actions and the expected outcome of its actions, in some trials the robot also shared the reasoning behind its decisions, its underlying logic. Participants viewed multiple Soldier-robot teams, and their assessments of the robots were compared.
The study found that regardless of the robot’s transparency in explaining its reasoning, the robot’s reliability was the ultimate determining factor in influencing the participants’ projections of the robot’s future reliability, trust in the robot and perceptions of the robot. That is, after participants witnessed an error, they continued to rate the robot’s reliability lower, even when the robot did not make any subsequent errors. While these evaluations slowly improved over time as long as the robot committed no further errors, participants’ confidence in their own assessments of the robot’s reliability remained lowered throughout the remainder of the trials, when compared to participants who never saw an error. Furthermore, participants who witnessed a robot error reported lower trust in the robot, when compared to those who never witnessed a robot error.
Increasing agent transparency was found to improve participants’ trust in the robot, but only when the robot was collecting or filtering information. This could indicate that sharing in-depth information may mitigate some of the effects of unreliable automation for specific tasks, Wright said. Additionally, participants rated the unreliable robot as less animate, likable, intelligent, and safe than the reliable robot.
“Earlier studies suggest that context matters in determining the usefulness of transparency information,” Wright said. “We need to better understand which tasks require more in-depth understanding of the agent’s reasoning, and how to discern what that depth would entail. Future research should explore ways to deliver transparency information based on the tasking requirements.”
The Latest on: Human-agent teaming
via Google News
The Latest on: Human-agent teaming
- Hang Ups: 10 Call Center Quality Monitoring Mistakes That You Need To Avoidon October 1, 2020 at 6:31 pm
By avoiding these 10 call center quality monitoring mistakes, you can improve your call center quality assurance scorecard & overall customer experience.
- 3 Smart Ways to Boost your Startup’s Customer Serviceon September 30, 2020 at 11:10 am
For your startup with a small customer support team, providing real-time support can be a challenge ... The focus should be on the interaction between a customer and a human agent, not a faceless ...
- WandaVision: How Scarlet Witch and Vision's children could change the MCUon September 30, 2020 at 9:04 am
working through a human agent. Unfortunately, the timing of this revelation couldn't have been worse for Wanda, who was still reeling from the corruption and destruction of her husband the Vision ...
- Innovative model improves Army human-agent teaming (image)on September 30, 2020 at 8:49 am
Army researchers develop a novel computational model for analyzing cognitive data that will be a game changer for the effectiveness of human-agent teams on the battlefield.
- Innovative model improves Army human-agent teamingon September 30, 2020 at 8:43 am
According to Bohannon, this effort will support the Next Generation Combat Vehicle via the Human-Autonomy Teaming, or HAT, Essential Research Program. One project under this effort is currently ...
- Innovative model improves Army human-agent teamingon September 30, 2020 at 8:05 am
According to Bohannon, this effort will support the Next Generation Combat Vehicle via the Human-Autonomy Teaming, or HAT, Essential Research Program.One project under this effort is currently ...
- Ignite 2020: Introducing Azure Communication Serviceson September 23, 2020 at 5:31 am
As Scott Van Vliet, CVP, Intelligent Communications at Azure noted, “The service that my team runs that powers ... in a bot and then moving to a human agent when more complex answers are ...
- Future autonomous machines may build trust through emotionon September 15, 2020 at 2:38 pm
“As an Army researcher, I am excited to contribute to this research as I believe it has the potential to greatly enhance human-agent teaming in the Army of the future.”The next steps for this ...
- Jamiahus Waltonon October 10, 2019 at 11:39 am
Generally, my research focuses on teams, team training, virtual collaboration, and computer-supported cooperative work. Recently, my work has focused on developing Intelligent Team Tutoring Systems ...
via Bing News