image: This is Richard Pak. view more
Credit: Richard Pak
The United States Air Force Office of Scientific Research awarded Clemson University a three-year, approximately $245,000 grant to determine what causes people to trust and rely on automation. The Air Force is reacting to two major trends: increased computerization and an aging population.
This research will seek to answer the following questions:
- Why do people continue to use unreliable automation?
- Why do people not use reliable automation?
- How do we get people to change their behavior about reliable and unreliable automation?
Our focus is to study how to make automated systems that people can trust and use. Automated systems permeate our lives and people seem to have a complicated relationship with automation.
Examples of everyday automation include: the spell-checker in email that highlights misspelled words; the assistant on cell-phone sites that help you pick the best phone and plan based on your usage; and the virtual assistant, Siri, that lives within the iPhone 4S. There are many other examples of everyday automated systems that millions of people encounter every day.
Our research will determine what makes automation trustworthy. Younger and older adults are likely to respond to automation in different ways. For example, younger adults may have less of a need for automation but older adults, because of normal age-related changes in memory and attention, are likely to need more help in understanding the myriad of choices available to them.
Choosing a cell phone plan and service is one example, but a more complex example is selecting a supplemental Medicare coverage or understanding when to take medications if one is diabetic. Healthcare is experiencing rapid changes, such as fewer doctors and less face time with healthcare providers, which means that patients may be interacting with web-based or mobile automated systems at home that will help diagnose or treat them.
Assuming these health aids are reliable, the optimal decision would be for the patient to use these aids. However, the decision to use a highly reliable automated system depends on whether or not we trust it or not.
Why we trust the automated system is the focus of our research. We will examine how the design of the automation (look and feel) affects a user's tendency to trust and, therefore, use automation.
Much of the existing research on automation has examined how engineering aspects of the automation, such as machine reliability, affect user trust. Only recently has attention been directed to psychological or social aspects of automation. As automation becomes more advanced and interactive, it will be treated as a team member or person. Our research will study this psychological aspect of automation and understand how people treat and use automation when it seems to behave and look like a human.
We must understand why some types of automation are trusted and used and why others are not, especially for an aging population. When we know the reasons, we can design automation that is trustworthy and can be relied upon.
Our first studies will examine the role of automation appearance on trustworthiness. We will solicit young adults (students) and adults aged 65 and older.
Related research: Clemson researcher: humanizing computer aids affects trust, dependence
Richard Pak is an associate professor in the Department of Psychology at Clemson University.