Project
Background
Artificial intelligence (AI) tools, like ChatGPT and others, are becoming more accessible to the general public. Due to this access, these tools are rapidly integrating into the education system. As current AI systems evolve, they have yet to be tailored toward kids aged 10-14; a generation of digital natives. From teachers’ perspectives, this age is crucial for a child’s development of critical thinking, creativity, and curiosity.
The research led to the creation of ThinkKee as a platform that facilitates critical thinking in kids through a series of activities, challenges, and guided learning experiences. It also empowers parents to make confident, informed decisions about their kid’s interaction with AI.
ACHIEVEMENT
I was honored to receive an SVA Alumni Scholarship for my work on this project.
ROLE
UI-UX Designer
FACULTY ADVISORS
Adriana Young, Elissa Ecker, Ida Benedetto
PROJECT
Design Thesis 2024
DATE
Sept 2023 - May 2024
INDUSTRY ADVISORS
Alexandra Risvang (Designer, DesignIt Studio)
Paul Amitai (Exec. Strategy Designer, Listen)
RESEARCH CERTIFICATION
Human Subjects Research – Institutional Review Board (IRB) Approval
My Role
-
Conducted online co-creation sessions with child-parent pairs and in-person interviews with teachers, psychologists, and counselors to deeply understand the needs and perspectives.
-
Synthesized research findings to develop a comprehensive design framework.
-
Developed a feasible product solution tailored to the intended audience and rigorously tested its effectiveness.
-
Created an online resource hub to educate parents and AI enthusiasts about relevant topics
-
Developed a customized GPT model with carefully crafted prompts to cater to the specific needs of under 10-14-year-old students.
Showcasing ThinkKee at the SVA Thesis Festival 2024!
Summary
Unlike conventional educational apps that may offer direct answers, ThinkKee introduces a layer of ‘friction’ while interacting with AI, designed to encourage users to explore solutions independently before receiving spoon-fed solutions. This approach makes the learning process more active, reflective, and engaging. The app features a dual onboarding process for both parents and children, ensuring parents are actively involved in establishing the learning goals, preferences, and safety settings for their children and fostering a collaborative learning environment.
This project also includes an online AI awareness guide for parents, consisting of simple steps to help evaluate a new AI tool and relevant sources of information, like current news articles, weblinks, and podcasts in the world of AI.
ThinkKee AI App
1. Shared Onboarding
The kid or parent/ guardian cannot register on the app individually. Both are necessary to set up the app. This ensures that the kid feels agency in using the app, while the parents are assured that the AI is suited well.
A 'General preferences' list and 'Accessibility preferences' list are provided in the onboarding to tailor the AI to the kid’s personality and accommodation.
2. Interactive & Informative homepage
Word-of-the-day, quote-of-the-day, News-around-us, and Question-of-the-day are the highlights to improve language skills and general knowledge.
These features on the homepage are provided to induce critical thinking of the kid
3. AI Chatroom
The AI provides hints, asks probing questions, and encourages exploration instead of offering immediate answers. It is safely made for kids to utilize.
ThinkKee provides relevant activities for the kid to perform without any spoonfed results. E.g. If the Kid asks 'Why do plants need sunlight?', ThinkKee will respond 'Let us do an activity to understand...'
4. Activities & Challenges
Creative challenges, Wordle, Quizz time, Creative prompt challenges and many more focused on logic and thinking
A gamification way to involve the kid in productive work
5. Profile with app usage stats and shared settings
Live dashboard to measure the activities and parents could keep track of the kid spending time on the app.
There is also another unique feature of the app that allows the parent or the kid to change the settings of the app only when both of them are present and with accepted consent. That way parents could not force something on their kid and the kid would not manipulate the AI to get results.
Thesis statement
How might we foster critical thinking in kids in K-12 education with the advent of Artificial Intelligence?
Gathering Insights
I have always been fond of teaching and interacting with students. This was me teaching science to a class of 30 students in 6th grade in Mumbai, India. Our conversations sparked my idea to create something useful for this generation.
Following were all the quotes by kids gathered after the conversation. My approach to gathering insights for the project was all about creating a friendly atmosphere. I had this plan in my mind, to keep things relaxed & unstructured. So, I just had casual chats with them and slipped in some spontaneous questions about AI, like ChatGPT and Midjourney.
Remarks from a few parents, teachers, child counselors, and education facilitators
Co-creation sessions
An hour-long online session with 4 pairs of a kid and their parent playing carefully designed activities to gauge parameters namely ways of thinking, communication clarity, convincing ability, and critical analysis of information. The activities were performed once with a participant set 1
(a human interviewer & kid) and then set 2 (An AI [ChatGPT v3.5] LLM & the kid).
Following were the 4 activities:
In addition to the above, extensive secondary research was involved with analogous technology studies, AI implications, extensions, the social impact side, and expanded discussions around theories of change, the value of time, and usability.
Design
Framework
The top 3 insights that informed the creation of the Design framework were as follows-
-
“I think civics subject is hard for me, so I tell ChatGPT - you are a social science teacher, you have to explain to a 6th-grade student about civics, do it in simple language please...”:
-
“I often get a lot of weird doubts like what is the loudest sound I can hear/ biggest word in the dictionary and search this on Google...”
-
“ I like reading...First I would write everything in my style, then go to my mom as she is an English teacher then some final things on ChatGPT”
Early
Explorations
A story-boarding exercise to evaluate prominent use cases of a kid using the AI. A common use-case scenario was the kid using the AI app to complete some science-related homework given in the school.
The initial sketches focused on the Chat feature of the app and ways to make the interaction with AI more kid-friendly, rather than just another social media chat room.
Mascot selection to embody the AI- ThinkKee. A fox figurine was created as a fox stands for being wise and is not an everyday tamed animal like a dog or a cat. It had a quirky interesting play to it.
Information architecture in a simplified format to lay out the app working and conditions to proceed. This was tested with participants and cocreators before the high-fidelity designs.
Several Iterations were made by testing the flow, and design with participants over the next 2 months. The testing was dine by eye tracking, live flow inputs via online meets.
Final Design
Feedback
This stage involved product MVP release with high-fi prototype testing and approving the final design version. A thorough list of suggestions and improvements for future versions were noted, but overall the product went well and participants were interested in using the product in reality.
The Onboarding screen 'General Preferences' questions were duly appreciated by the stakeholders and realized the implications of those questions for setting the AI personality.
For the homepage, 2 important questions were raised-
-
How often will the homepage refresh?
-
Can the AI respond to the mood of the kid, if they are cranky or tearful and provide things accordingly?
There were a few other important points raised.
-
How often will the ThinkKee AI disagree to provide the solution? Current GPTs could be easily tricked.
-
The Activities could be expanded to Digital Detox challenge, Movement inducing games which are necessary to reduce screentime
-
The selection of learning styles under Settings followed Mihaly's Flow concept of easy, challenging, and adaptive learning controls. It acted as an important parameter for different kinds of children.
-
Selecting different genres of interest under Settings paved the way for knowledge exploration, without any repetition or same subject information.
Project
Reflections
-
Risk of the project becoming obsolete: The reason was newer, more sophisticated AI tools emerged. To mitigate this, I focused on creating a framework that could adapt to future AI developments, rather than relying on specific tools or technologies.
-
Working with children aged 10-14: It presented unique challenges due to their underage status and the limited availability of AI tools specifically designed for their age group.
-
Convincing parents was a significant hurdle: As many were unfamiliar with AI technology and its potential benefits, addressing their concerns about children's overreliance on AI and the need for independent thought was crucial.
-
Additional factors: Ethical concerns, technical challenges, impact assessment, and sustainability had to be considered throughout the design and development process.