Building Relationships

Learning through Reflection upon Experience

Torsten HardießKönigswinter (GER), December 2018 - Torsten Hardieß has investigated the issue of how "Alexa, chatbots, and similar tools individualize learning and improve transfer". The psychologist and his team are undertaking research into how human-machine interactions can be used successfully in the development of personnel and their individual personalities. He will speak at the LEARNTEC Congress, 31 January 2019 at 10:45.

Is it possible today to communicate with Alexa in regard to a complex learning plan, or does an Alexa manual offer little more than standard, general information?

Torsten Hardieß: I think we have to take a look at this question from a more differentiated perspective.

We are living in an era of big data, artificial intelligence, ever-faster computers, calculations in the cloud, machine-to-machine communication, and information-retrieval systems. In this context, training people with nothing but factual knowledge to become experts, such as lawyers or bank consultants, is not particularly future oriented.

Our goal is to support people in developing skills and abilities that are crucial to effective human interaction. I’m referring here to key competences such as being able to deal with conflicts, the capacity for empathy, or the ability to manifest leadership behaviors. From our point of view, these social skills are the key to a successful professional life and companies that flourish. This is why we’re asking ourselves how digital systems can contribute to this.


What are the learning needs for which Alexa, chatbots, and co. are suitable?

Torsten Hardieß: Videos that build on one another can easily convey factual knowledge and procedures linearly, and quizzes can then be used to measure learning success. When it comes to social skills, though, knowledge is often of little help. People who, e.g., correctly answer questions about the percentage of communication that takes place unconsciously after looking through the module on the "iceberg model" are very rarely less aggressive five minutes later when meeting a colleague in the corridor who has been getting on their nerves for weeks.

In contrast, we use "conversational interfaces" to stimulate user reflection; to simulate conversational situations; or, through memories and everyday life challenges, to motivate questioning of a behavioral pattern toward the goal of changing it. The interface, whether it’s text based or language based, assumes a role as coach.


How and through which functions does the improved transfer come about?

Torsten Hardieß: When virtual assistants are employed, a new measurement category comes into play. The central issue is no longer the user experience, in which IT is scrutinized for speed, accuracy, or the number of mistakes, but of how to establish a relationship between the virtual assistant and the human being. Thus, a lot of significance is attached to measuring the connection or trust between person and machine.

Numerous studies indicate how much people "humanize" objects and machines and construct relationships with them. The important thing for us is optimizing this experience for the users.

The interaction via text and language plays a central role in this process. After all, this is how we humans are used to communicating. If the assistance systems also have a personality - with rough edges and flaws - just like people - it results in exciting interaction possibilities.

Tasks delegated by this "personality" are more likely to be performed than something presented in an e-mail from an anonymous system. A relationship is also formed through the bot’s checking in with me regularly to ask what I have learned or, as a real coach figures out, what prevented me from carrying out the task. When metaphors, challenges, and content are combined into an exciting story, possibly supported by gamification elements, interactive, fun learning experiences are created.


Will people who resist the use of virtual assistants fall behind in terms of learning?

Torsten Hardieß: Of course there are still people today who use candles. Actually, assistants are already in use in cars, in every mobile device, and even in modern food processors.

We know that people don’t learn from experience, but rather by reflecting on their experience. This is why I use my virtual assistant in the car between customer appointments to reflect on what I could have done better.

Whenever I want, the bot gives me the freedom to be stimulated with the right questions while I’m reflecting.


What technical developments do you expect to see in bots and virtual assistants over the next few years?

Torsten Hardieß: The understanding of language will improve, and there are already some exciting approaches to making language systems sound much more human. This will also contribute to building trust.

Furthermore, assistance systems’ ability to understand language content will also improve. Our own research, for example, is focused on teaching bots the ability to understand people empathetically.