The fundamental premise of the project is based on the therapeutic potential of laughter. It has been widely accepted that laughing often can enhance lifespan and overall health by alleviating stress, amplifying the immune system, easing pain, and spreading happiness to those around us. Laughter, a contagious phenomenon, fosters positive emotions and a sense of community. This project delves into whether artificially generated laughter can produce a contagious effect similar to that of natural laughter and further explores the potential and pathways for how artificially generated laughter might be applied to human life and industries in the future.

my role was

- 100% of Coding

- 70% of UI Design

- 40% of Research

tool

- Figma

- Html / Css / Javascript

i wokred with

- Wenqi Yan

year

- MARCH 2023 ~ APRIL 2023

class

- IXD 636 - Artificial Intelligence

Problem Statement &
project Goal

We embarked on this project with a central question in mind: 'How might we use AI to promote laughter among humans?' The target users are individuals seeking to harness the benefits of laughter. The ultimate goal of this project was to explore the potential of AI in generating humor and fostering laughter, and to compare its effectiveness with traditional, human-led humor generation.

solution

LET'S MEET SELINA

The solution we devised was an AI character named Selina, the AI laughter therapist. This character allowed users to interact with an AI-generated voice. The laughter yoga technique we adopted involves prolonged voluntary laughter and is believed to provide similar benefits as spontaneous laughter. Additionally, we utilized the defamiliarization strategy, which unconventionally presents familiar topics to generate unexpected humor. These were incorporated into our AI to create a distinct and engaging user experience.

how it works

1. Standby Mode: This is the initial state of the AI when it's ready to interact with the user but isn't actively engaged yet.
‍
2. Eye Contact: The AI initiates interaction by making virtual 'eye contact' with the user, acknowledging their presence and readiness to engage.
‍
3. Self-introduction: The AI introduces itself to the user, creating a friendly and personable atmosphere.
‍
4. Start laughter: The AI seeks the user's consent or willingness to proceed with the session, ensuring they are comfortable and ready to participate.
‍
5. Imitating AI laughter: Upon receiving a positive response, the AI generates laughter, prompting the user to mimic or follow the laughter.
‍
6. Laughter Score: The AI uses the user's laughter volume data to calculate a 'laughter score.' This score could provide feedback to the user on their engagement and responsiveness to the laughter therapy session.

Two key
components

We incorporated the principles of defamiliarization and *laughter yoga technique into the AI laughter therapist. These could make AI a tool that can effectively promote and induce laughter in a therapeutic context.

*Laughter yoga technique
Laughter Yoga is a practice aimed at improving physical and mental health through laughter. It was first developed by Dr. Madan Kataria in 1995 in India. This technique capitalizes on the physical and emotional benefits that come from spontaneous laughter, which can start with fake laughter and evolve into genuine laughter through interaction, eye contact, and group activities. The purpose of Laughter Yoga is to offer health benefits such as stress reduction, improvement in mood, and strengthening of the immune system. It aims to facilitate these benefits by inducing laughter in situations where it might not naturally occur.

How we got to the solution

first protype

We were curious about whether artificial intelligence truly had the potential to make people laugh. Therefore, we created the first prototype of Selina, an AI laughter therapist, for testing purposes. The process for the first prototype is as follows: After passing through the splash screen and pressing the start button, a page with Selina appears. When users mimic Selina's laughter on this page, a laughter score is displayed at the end.

How it works

1. Splash screen

2. Start page

3. Laugh with Selina

4. Score

How it works

1. Splash screen

2. Start page

3. Laugh with Selina

4. Score

How it works

1. Splash screen

2. Start page

3. Laugh with Selina

4. Score

How it works

1. Splash screen

2. Start page

3. Laugh with Selina

4. Score

User Testing &
learning

We conducted tests showing participants the first prototype of our AI laughter therapist, Selina, and then videos of humans laughing. The results were intriguing: more participants laughed at the human videos, yet many still found amusement in Selina's laughter, albeit to a lesser extent than in the human videos. From this experiment, we found the potential of AI to induce laughter among people.

demonstration &
Feedbacks

During class, we demonstrated the prototype of Selina, developed based on user testing. We received comments that the kitschy and childlike UI distracts the users and received feedback to rethink our project toward improving User Trust, Familiarity, and Engagement.

improvements
from feedbacks

We iteratively improved our solution based on the learnings from user testing and feedback to enhance our AI system progressively.

1) Emulating Life: Enhancing User Engagement in Standby Mode

[Before]

[After]

In the early stages, our AI, Selina, remained static in standby mode. Users struggled to perceive if Selina was ready to interact. So, we made Selina blink her eye and subtly move her face, mimicking natural human behaviors. As a result, the user interaction in standby mode significantly improved over the initial design.

2) Humanizing Interaction: Enhancing User Trust and Familiarity

We made Selina introduce herself by saying her name when making eye contact with the user. This is because as Selina, an AI character, provides such human-like interactions, users can perceive the AI as more 'human' and familiar. When AI converses and responds like a human, it allows users to feel like they are interacting with a natural person, not a machine or a computer. This fosters a more comfortable response from users and enhances their trust in the assistance provided by the AI.

3) Visual Feedback: Enhancing User Confidence in AI Interaction

When Selina emits the 'HA' sound for laughter, the user is meant to mimic her laughter. Initially, the screen had no visual indication when users followed Selina's laughter. As a result, users were unsure if the AI was detecting their laughter. We addressed this by adding a color bar on the screen when users laughed, visually signaling that the AI was detecting their laughter.

4) design: less is more

[Before]

[After]

We redesigned the previous kitschy and childlike UI by minimizing colors, fonts and other design elements to let users focus on Selina's function as a laughter therapist. The overhaul aims to faithfully execute Selina's capabilities and keep the users' attention on its features.

result

We displayed Selina's prototype at a student exhibition. As a result, we were able to confirm that 70% of students laughed with Selina.

Future Applications of
AI-Generated Laughter

AI's ability to induce laughter is crucial for several reasons. Laughter is a universal language of connection, known to have profound effects on our physical and emotional well-being. It reduces stress, boosts the immune system, and releases endorphins, which are the body’s natural feel-good chemicals. AI that can generate laughter has the potential to contribute to mental health, foster social connections, and improve quality of life.

Although this project is in the prototype phase of testing the feasibility of artificial intelligence in eliciting human laughter, its significance and potential are profound. If AI can develop the ability to interact with humans through a sense of humor, the applications could extend across various societal sectors. Therefore, the lessons and feedback from this stage could have a significant impact as the project moves towards a comprehensive product development phase.
‍
In terms of future applications, AI-generated humor could be integrated into various industries:

1) Mental Health: Humor can help reduce stress and anxiety. AI-based therapeutic tools or applications could utilize humor to support users' mental health.

2) Entertainment: Humor plays a role in enhancing the enjoyment of everyday life. If AI can understand and generate humor, it could be used as an auxiliary tool for creators such as novelists, scriptwriters, and advertisers. This can stimulate the development of creative content and create new forms of entertainment. AI-generated humor can be applied in various fields such as advertising, customer service, and the entertainment industry, creating new value across industries.

3) Education: Educational software can be designed to incorporate humor, making learning more interesting and enjoyable for students, thereby improving attention and retention rates.

What I've learned

‍1. I enhanced my coding skills throughout this project, focusing on design modifications for better user interaction. Exploring the confluence of technology, psychology, and therapy with Selina's development was an enlightening experience. The project gave me a versatile set of skills and was an enriching technical journey.
‍
2. Through this project, I learned a lot about how to make people feel more familiar with and trust AI, and this knowledge will be a great asset in future projects.Someday in my UX career, I believe I will have the opportunity to work on a similar project related to AI assistants.

SelIna: the A.I. Laughter Therapist

How we got to the solution

1) Emulating Life: Enhancing User Engagement in Standby Mode

2) Humanizing Interaction: Enhancing User Trust and Familiarity

3) Visual Feedback: Enhancing User Confidence in AI Interaction

4) design: less is more