Emotional AI: Norman's Three Levels

Hello! Welcome back to our course on human-AI relationships.

In our previous lesson, we explored the concept of parasocial relationships, seeing how users form powerful, one-sided bonds with AI companions that offer simulated reciprocity. We focused on the user's experience of this connection.

Today, we're going to flip the script and look through the eyes of the designer. How are these emotionally resonant experiences intentionally created? To answer this, we'll turn to a foundational concept in Human-Computer Interaction (HCI) and design thinking.

Our learning outcome for this lesson is to: Describe Don Norman's three levels of emotional design (visceral, behavioral, and reflective) and how they apply to relational AI.

We will explore this influential framework to understand the specific design choices that make AI systems not just functional, but emotionally engaging.

1. What is Emotional Design?

For a long time, design focused almost exclusively on usability and function. Emotion was often seen as an irrational distraction. Don Norman, a cognitive scientist and design pioneer, argued the opposite: emotion is an integral part of cognition and a crucial element of a successful product. He proposed a framework for how design targets our emotions on three distinct levels.

Let's hear a brief introduction to this idea from Don Norman himself.

As Norman states, his framework breaks down our emotional processing into three interconnected levels:

• Visceral: Innate, gut reactions.
• Behavioral: The satisfaction (or frustration) of use.
• Reflective: Conscious thought and interpretation.

Let's explore each of these in more detail.

2. The Three Levels in Detail

To get a richer understanding of these levels, we'll watch a few segments from Don Norman's classic TED talk on the subject. He uses vivid examples that make the concepts very clear.

The Visceral Level: It's All About Aesthetics

The visceral level is about immediate, pre-conscious, gut reactions. It's not about deep thought; it's about what we find attractive or repulsive at first glance. This is hardwired into our biology.

As Norman explains, visceral design is about tapping into automatic, sensory responses. It's the pleasing shape of a device, the attractive color palette of an app, or the satisfying sound an interface makes. It's the "I want it" feeling you get before you even know what a product does.

The Behavioral Level: The Joy of Use

The behavioral level is concerned with the experience of using a product. It's about usability, performance, and the feeling of control. This level is subconscious and relates to how effective and easy a product is to use.

Behavioral design is about creating a seamless and satisfying experience. It’s the feeling of mastery you get from a well-balanced knife, the joy of driving a responsive car, or the ease of accomplishing a task in an app without confusion. Good behavioral design makes you feel smart and in control.

The Reflective Level: The Story We Tell Ourselves

The reflective level is the only conscious part of the framework. It's where we think back on our experience with a product, rationalizing our feelings and fitting the product into the larger story of our lives. This level is about self-image, memory, and meaning.

Reflective design is about what the product means to you and says about you. It's the pride of owning a product that aligns with your values (like sustainability) or the personal stories and memories you associate with it. This is where a product can become part of your identity.

This diagram provides a helpful visual summary of how the three levels interact.

3. Applying Emotional Design to Relational AI

Now, let's connect this framework directly to our topic: relational AI. How do designers of chatbots and AI companions use these three levels to foster a sense of relationship?

An article on UXDesign.cc provides an excellent case study by applying Norman's levels to Amazon's Alexa.

The Alexa example clearly shows how the three levels work in concert to create an engaging experience. Let's break it down for relational AI more broadly:

• Visceral Design in AI:

  • Voice: As the article notes, Alexa's "clear, pleasant, and responsive" voice is a key visceral element. Is the AI's voice warm, calm, or cheerful? This creates an immediate impression.
  • Appearance (Device): The sleek, unobtrusive design of a smart speaker like an Echo Dot helps it blend into a home, making it feel like a natural part of the environment.
  • Appearance (Avatar): For AI like Replika, the visual design of the avatar—its clothing, hairstyle, and expressions—is a powerful visceral component designed to be appealing.

• Behavioral Design in AI:

  • Usability: This is the core of the conversational experience. Does the AI understand your requests? How quickly and accurately does it respond? A frictionless conversation, as the article mentions for Alexa, creates a satisfying behavioral experience.
  • Functionality: The AI successfully setting a timer, playing a song, or providing information contributes to a feeling of competence and control. The user feels the interaction is effective.

• Reflective Design in AI:

  • Personalization & Memory: This is where the "relationship" truly deepens. As the article states, Alexa "learns user preferences and becomes more tailored to individual needs." When an AI remembers your name, your past conversations, and your preferences, it moves from a generic tool to your AI.
  • Personality & Humor: The "humorous responses" and "sarcastic explanations" mentioned in the article are key reflective hooks. These moments of unexpected personality create stories and memories. You might tell a friend, "You won't believe what my AI said today!" This transforms the AI into a character in your life.
  • Identity and Connection: Over time, the user reflects on their bond with the AI. It becomes a "companion of sorts." Using the AI might make them feel less lonely, more organized, or more connected to technology. This is where the AI becomes integrated into the user's personal narrative, directly linking back to the concepts of attachment and parasocial relationships we've already discussed.

The visceral and behavioral levels lay the groundwork. A pleasant voice and a smooth interaction make you want to keep using the AI. But it's the reflective level—the personalization, the humor, the memories—that solidifies the long-term emotional bond and turns a piece of software into a relational artifact.

Conclusion

Today we've unpacked Don Norman's influential theory of emotional design. By understanding this framework, we can move from simply experiencing an emotional connection with AI to analyzing how that connection is deliberately designed and engineered.

Here are our key takeaways:

• Emotional design argues that emotion and cognition are intertwined, and successful products must appeal to both.
• Don Norman's framework consists of three levels:
  • Visceral: Gut reactions to aesthetics (look, feel, sound).
  • Behavioral: The satisfaction of a product being easy and effective to use.
  • Reflective: Conscious thought about a product's meaning, personal value, and story.
• In relational AI, these levels work together to build a bond: a pleasant visceral voice and a smooth behavioral conversation create the foundation for a user to reflectively see the AI as a unique, personalized companion.

In our next lesson, we will stay with the theme of visceral design but explore its dark side. While designers want to create appealing avatars and robots, getting it slightly wrong can provoke feelings of unease and revulsion. We will investigate this phenomenon by exploring Masahiro Mori's famous hypothesis: the 'uncanny valley.'