Emotional design in human centered AI

Emotional design in human centered AI

In order to touch consumers, data needs to find the heart.


Hardly any other buzzword has been thrown around more profusely this year than “artificial intelligence”. Virtually every chatbot, no matter how rule-based, has been adorned with the magic attribute “AI” just to look innovative and stir up expectations and desires. Look a little closer and you will realize that most bots follow a calculable logic that only allows for some or even no degree of autonomy. A more or less complex decision tree mechanic that engages a little bit more in a dialogue with the users in their habitual eco-systems – Granted! But intelligent?

The magic that spellbinds so many is based on self-learning and above all reactive algorithms that have to be preceded first of all by a learning process. But what is it we do with this data? Do we understand our users better and then respond to them in an adequate fashion? Do the users feel at ease, taken care of and appreciated? Or even better: Do they identify with the brand thanks to the dialogue that took place and thus follow the brand on account of the authentic personality being presented? We need emotionally intelligent interactions via interfaces that can be integrated in the lives of our users. Seamlessness means that we consider all sensory channels and learn how to communicate in a way that is context-relevant.

Data vs. emotions

Emotions & personality, in addition to data-driven interactions, are important for intelligent user experiences. What we should be thinking about when considering a comprehensive experience with a product or service is a personality that is rooted in the brand’s core. Whoever wants to create relevance in times of oversaturated markets should put some thought into how the brand behaves, feels like and what it stands for. It’s these properties that are needed for a “unique” character that stands out in clear contrast to other brands.

At the most fundamental level, of course, customer’s needs have to be met and their problem solved. This can be achieved through adequate qualitative research in the field of user-centered research on prototypes elaborated in various degrees. Exactly here you can find the right elements that require fine-tuning and thus create the appropriate form of address, since tonality can be found in everything - in language, design and the system’s feedback.

Think of a product you truly love. A product you would recommend to your friends. And then figure out how these emotions could be evoked in you in the first place.

Emotional product design

Emotions are very powerful and influence our perception, our decision-making abilities and memory. The more emotional an experience, the stronger our memory. According to Aaron Walter’s theory, a product should be functional, reliable and usable (in this order). Emotional design is the crucial level that is added to a product as soon as pure functionality has been ensured.

“Emotional Design should never interfere with usability, functionality, or reliability.” (Walter)

Don Norman talks about Visceral Design, i.e. a design that evokes instinctive or maybe even subconscious responses. Imagine your favorite tech author recommends the latest software application to optimize your expenditures. The impact of this recommendation differs from the impact of a paid online commercial even though we are talking about the same software application.

The next level is Behavioral Design. This is design responsible for how a product really feels like – how it behaves. In the digital world this might be micro-interactions, the tone of voice and the design’s aesthetics.

The third level is the Reflective Design. This means the impulsive pondering of products as soon as a trigger kicks in. Picture yourself riding on a train and your smartphone battery is drained empty. Exactly now you are bound to realize how great Instagram and bretheren would be to kill some time. At this point a good service has already managed to make it into the Relevant Set and thus create an emotional bond with its users. If you now also bear in mind the expectations of a seamless experience and modern interaction, the thought insinuates itself that the challenges grow exponentially with additional channels, such as voice interfaces or messenger services.

How can you behaviorally condition users in such a way that they understand that following a learned interaction will bring them a kind of a reward that will fulfill their expectations? This form of gamification is responsible for at least a mild form of addiction in the user. Put more positively, we are talking about loyalty here.

Artificial personality

Our social and cultural background influences all of us. Naturally the manner in which we have learned to respond to our fellow humans’ impulses from early childhood on, plays a role. This has made us who we are and puts us into a complex framework of interpersonal rules. Everything else we have to learn to communicate with each other is based on language, mimic, gesture and behavioral codices. We live in a constructionist world and when dealing with people, expect to find models that correspond to those we have learned. Fact is, we also have to learn foreign languages from scratch and “feel” them before we can communicate smoothly. The same holds true for computer systems, but here we like to use metaphors, mental models and skeuomorphisms to break down the barrier between man and machine. Think about the desktop metaphor that has been the accepted model for many decades now.

What else can be more immediate than established patterns from your own reality? The intuitive access opens up only when I don’t have to learn a system language from scratch but when this language adapts to my expectations. This means, you prefer human interaction to machine interaction. This is why products with a personality and a sense of humor help users to identify with a product or a service. The result: humanizing computers – anthropomorphic computers. With the clear goal of winning users’ trust by communicating eye to eye with them.

Machines or systems assume a context-related personality. This process has been going on for much longer than it seems at first glance. Since the 50s people have been working on voice-recognition programs and since 2000 this field has arrived at a new level. For instance, Siri started formulating witty answers in 2011. This marks the birth of voice assistance and at the same time personality in the automated man-machine interaction. Take, for example, brands, services, products … etc. that are represented by personae that reflect a certain tonality in the interaction. Amazon Echo, Google Home and other Voice User Interfaces are advancing this field and the system will learn to become more human in the future (machine learning). Intentions will be anticipated instead of just responded to by aggravating requests such as “Please press 3 for questions about assembly!“

Also, users’ expectations are adapting to the new possibilities. A Google and Amazon research study reveals: Whenever machines start a dialogue with us – via Conversational User Interfaces or Voice User Interfaces – we, the users, expect genuinely emotional relationships. The Chatbot or Voice assistant turns into a “personal assistant”, up to the point of being “best buddy” as in the movie “Her”.

The challenges in developing such systems are quite complex. No set User Flows – but more contextual conversations. Natural languages do not follow a top-down information architecture. There is no rigid decision tree that only responds to key words. The system must be able to ask the right questions! The challenge any designer has to face is: to understand the user’s context!

Here you can see a clear evolution of a new interaction paradigm emerging, the de-materialization of the human-computer interface that is! What started with vacuum tube computers, punch cards and Command Line Interfaces and today’s ubiquitous General User Interface, is now moving towards a Natural User Interface with touch, voice or gesture operation that is morphing from a tangible to an intangible, non-corporeal, companion that is integrated seamlessly into our daily lives and can be operated via well-known natural interactions. The results in systems like Digital Agents that respond so intuitively and naturally that they are able to create an emotional bond with the user and convey personality by reacting intelligently and anticipation.

In the future, systems will adapt to human needs so naturally that the use of technology will no longer be a barrier.