This is the last of my current projects and hopefully provides an illustration of how new research directions build off of existing research, with new questions and new goals.

The intent of this project is to develop Empathic Tutorial Bots for Student Populations.

In going over earlier posts you probably have noted the issue of dependency/addiction in relationship to the digital, whether it be Korean gaming addiction, social media addiction, Second Life/virtual world dependency-escapism, etc.  All have a common theme with the passive place of the user and the manipulation by various external forces.

However both Miller’s group (the Why We Post project) and our work indicate that we have to start looking at the user as an active rather than passive agent.  When you make this conceptual shift, you also end up looking at issues of dependency differently.  Ironically, seeing users as active agents suggests that ‘removing’ a dependency can be more of a struggle than with the passive agent model.  With the active agent model you have to provide a viable alternative to the behavior they want to control.  This means that if you want to get someone to shop checking their Instagram or WhatsApp account every 30 seconds, you have to come up with something more interesting, more attractive, or more emotionally rewarding.

Research on dependency-addiction in various forms has consistently shown that emotional engagement is critical to long-term success.  The user has to care about “X” enough that they are willing to modify their behavior.  This is the basis of Alcoholics Anonymous groups, support groups, support networks, etc.—all use the process of building emotional engagement with others to encourage the behavior you want to modify.  Once this happens, getting (and keeping) the group’s approval for ‘good behavior’ becomes a major motivator for behavioral change.

Effective learning also involves a strong social-emotional component.  The more emotionally engaged, the more effort is put into the learning process.  This is why one-on-one tutoring and small group training is usually more effective than large class instruction.  But traditionally issues of time, cost, and availability all work against the application of this technique on a large scale.

The Bot/AI field has seen a number of different designs intended to address these issues, but Hanson Robotic’s Professor Einstein Tutorial Bot (https://professoreinstein.com/) is arguably the first successful iteration.  By combining the Bot/Tutor (verbal) with the App [Stein-o-Matic](tutoring materials) they have built an automated tutorial Bot.  By virtue of its appearance, movements and voice it is intended to build an emotional relationship with the user/student.  The intent is that the more the student emotionally engages with the Bot, the more they will pay attention to the tutoring process, the more effort they will put into working through the tutorials, and the more they will retain the material.

Providing emotionally engaging, individualized, on-demand tutoring and behavioral coaching is the goal of our project.  We are developing a Tutorial Bot which by virtue of appearance, movement and sound will encourage the user to build an emotional relationship with the Bot.  The Bot is intended to be held and thus through physical contact build stronger emotional bonds with the user.

In our version the Bot acts both as a Learning Tutor and also as a “Life Coach”.  The “Life Coach” function is driven by physiological monitoring to record and track stress levels.  One of the main goals of the monitoring is to assist in the mitigation of negative behaviors related to stress and social media dependency.  At the holographic level, Gatebox’s ‘Maid’ AI system (https://gatebox.ai/) apparently has some of the same goals.

The Empathic Bot is intended to provide educational and behavioral support to university students to both assist their educational success and reinforce positive behaviors to support their emotional health.