Man stands in front of screen to speak about his research

Meet Torsten Maier


As part of the Technology and Human Research in Engineering Design (THRED) Group at Penn State, Torsten Maier, a Ph.D. student in the Penn State Harold and Inge Marcus Department of Industrial and Manufacturing Engineering, actively studies the effects of AI on human trust to understand the human and AI relationship.

"I graduated as a mechanical engineer from the University of Evansville," Maier said. "I chose Penn State because when talking to other graduate schools around the country, no other faculty spent nearly the same amount of time talking to me about their research and projects. They worked to find where I could fit in based on my interests and background."

The THRED Group at Penn State is a team of interdisciplinary researchers studying people, products and processes. Among them are Chris McComb and Jessica Menold, both assistant professors for the School of Engineering Design, Technology, and Professional Programs. McComb and Menold serve as co-directors.

McComb explained that Maier's work started when they identified a shortcoming in the design of cognitive assistants. McComb said there wasn't a precise definition of the design space for cognitive assistants, which then allowed designers to develop new cognitive assistants ad hoc.

"It was kind of like shining a flashlight into a dark room," McComb said. "We hope that his work will help inform the design of the next-generation of cognitive assistants."



Maier's most recent study, "Assessing the Impact of Cognitive Assistants on Mental Workload in Simple Tasks," took a look at the intersection of human attention span, or cognitive cost of doing a task, and human performance.

Unexpectedly, Maier received a robust verbal response from participants on trust. Maier explained that this understanding of the human and AI relationship could help inform future AI policies and provide a better AI–human experience.

Popularly called the Wisconsin Card Sorting Test, several cards were presented to the participant for them to match sans instructions. If used, the cognitive assistant would make suggestions for the next move, which was designed to aid performance and reduce frustration. 

"The cognitive assistant in the test was the research team," Maier said. "We sat behind a one-way mirror and fed answers to a Bluetooth speaker. It was like the Wizard of Oz where there's a big head in the room, but behind the curtain is just a man pulling strings. The answer given by the cognitive assistant was always right."

His study found that participants using the cognitive assistant for assistance were able to perform substantially better than those that did not. After the participants finished the game, the team interviewed them on their experience. The participants were eager to share their thoughts on trust despite there being no trust-based questions.

"The participants told us that they didn't believe the cognitive assistant was going to be right and that it took them multiple rounds of asking for help before they began to trust it," Maier said. "They figured out that cognitive assistant would have the right answer when they got stuck, which ultimately made the test easier for them."

The results based on trust showcase the critical ways in which cognitive assistant design can alter a person's feelings towards it. 

Maier explained that affective trust, or emotional trust, can be done through changing the voice of a cognitive assistant from male to female, or vice versa. There is also analogical trust, otherwise known as trust through trial and error. Lastly, designers must consider analytical trust, which is when someone trusts based on seeing values and statistics. There needs to be a certain level of manipulation on each of these categories to have a successful cognitive assistant.



Maier chose this line of work because he believes that the lack of standardization and classification among cognitive assistants raises questions towards their regulation and governing, particularly in healthcare, where vulnerable populations, like the disabled or elderly, may rely on them to complete day to day tasks. 

In addition to his experiments, Maier is working on the creation of a formal ontology that could significantly assist in the regularization of cognitive assistants to create adequate policies for this rapidly budding technology.

Maier explained that unlike a taxonomy, such as the animal kingdom, an ontology is not mutually exclusive and is categories of something. He noted that the THRED Group identified learning, intelligence, communication and autonomy as four categories of a cognitive assistant, but there can be an expansion with additional classes and subclasses.

"Industrial engineering is widespread, you can pick what you want to do, and I think that is pretty awesome," Maier said. 

Maier has presented his work in Anaheim, California, at the American Society of Mechanical Engineers 2019 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference and in the Netherlands at the 2019 International Conference on Engineering Design.

The student spotlight series by the Penn State Harold and Inge Marcus Department of Industrial and Manufacturing Engineering aims to highlight innovators, makers and those that personify engineering excellence in their academic studies. The department currently has 90 doctoral students, 59 master's students and 436 undergraduate students. In addition, the department hosts 31 full-time and courtesy faculty members. Established in 1908, the department is home to the first industrial engineering program in the world and has made a name for itself in the engineering industry through its storied tradition of unparalleled excellence and innovation in research, education and outreach.

Cognitive assistants, a popular form of artificial intelligence (AI) like Apple's Siri or Amazon's Alexa, are a rather useful tool in today's society. An individual can use a cognitive assistant to check the weather, determine travel times, send a text or call a friend. Despite the advances in this technology, there are still unopened doors for research regarding the impact that cognitive assistants have on human lives and how much we believe them to be correct.



Home of the first established industrial engineering program in the world, the Harold and Inge Marcus Department of Industrial and Manufacturing Engineering (IME) at Penn State has made a name for itself in the engineering industry through its storied tradition of unparalleled excellence and innovation in research, education, and outreach.

We are Innovators. We are Makers. We are Excellence in Engineering. We are Penn State IME.

The Harold and Inge Marcus Department of
Industrial and Manufacturing Engineering

310 Leonhard Building

The Pennsylvania State University

University Park, PA 16802-4400

Phone: 814-865-7601

FAX: 814-863-4745