Redesigning Khan Academy: Improving Accessibility for Learners
Existing digital learning experiences can create additional cognitive and visual load for autistic learners. How can we resolve this, within a top platform such as Khan Academy?
Client:
Khan Academy
Category:
UI/UX Design, Product Design
My Role:
UI/UX Designer, Product Designer, UI/UX Researcher · Team of 4
Children with autism often struggle to engage with reading experiences that do not accommodate for variations in communication style, attention span, and sensory sensitivity.
When the COVID-19 pandemic hit in 2021, children on the ASD spectrum were left behind to adapt to a form of learning that turned upside down.
The Centers for Disease Control and Prevention reports that 1 in 54 children have been identified with autism spectrum disorder by age 8. For those on the spectrum, they face difficulty communicating and interacting with others.
The global e-learning market was valued at $486.41 billion in 2025 and expected to reach $740.46 billion in 2032. It's estimated that by 2029, about 995.9 million people will utilize online learning platforms in the U.S.
As e-learning becomes more and more prominent and equivocal to physical classrooms, e-learning businesses need to make their services more accessible to autistic students, as their needs are overlooked in standard app development.
Because of their special needs, current e-learning platforms make it a challenge for them to process information and progress at their own pace, often leading to disengagement and limited learning outcomes.
👶 When learning materials are too difficulty, boring, or doesn’t align with a child's interest, children may easily become frustrated, disengaged, or depressed.
💻 Tools that aren’t designed for ASD often fail to support learning comprehension without extra guidance.
📉 Misaligned experiences can lead to reduced engagement and learning outcomes.
Our Research
Recent research provides significant insights into how children with Autism Spectrum Disorder (ASD) perceive and associate colors compared to typically developing (TD) peers. This research highlights atypical preferences and a unique strength in binding specific color-shape associations.
Feature | Typically Developing (TD) Children | Children with ASD / High Autistic Traits |
|---|---|---|
Top Preference | Red | Red / Green |
Least Preferred | Brown / Pink (boys) | Yellow / Brown (under age 11) |
Sensory Reaction | High luminance (Yellow) is bearable | High luminance is sensory-overloading |
Cognitive Binding | Flexible color-shape associations | Rigid/Stronger color-shape associations |
Implications for Design and Clinical Settings
Sensory Management: Reducing the use of high-luminance colors like yellow in classrooms or clinical tools can help prevent sensory overload for children with ASD.
Leveraging Associations: Since color-shape binding is stronger in those with high autistic traits, using consistent color-coded systems (e.g., always using a red circle for a specific task) can be a highly effective way to facilitate learning and memory.
Visual Aids: Using colored filters or overlays has been shown to provide perceptual benefits and improve the detection of social cues in some individuals with ASD.
The Design Process
We first looked at Existing General Learning Apps.
And then at ASD Specific Apps.
General reading and early learning apps like Khan Academy Kids and ABCMouse are successful because they provide engaging visuals, rewards, and structured progression systems, helping maintain attention and motivating continued use. However, being designed for neurotypical children they rely on fixed pacing, dense visual layouts, bright colored animations, and less flexible interaction models.
Apps designed for children with ASD such as Proloquo focus on communication support, behavioral learning, and structured skill development. These tools are valuable because they emphasize simplicity, structure, and individualized support.
The Good, The Bad, and Ugly
The Solution
4 user interviews later, with distinguished founders/professors in the field of ASD/child therapy…..
Came the current prototype.
Problem: Autistic children are more prone to lose attention to the task at hand, which made online learning a difficult shift during the pandemic when no teachers were present to refocus them.
Solution: Our engagement prompts utilize gently wiggling animal mascots and eventually audio, that activate after a certain time of inactivity.
Problem: Current e-learning platforms do not universally offer a variety of color palettes to be applied throughout the entire screen. They are typically limited to Light or Dark mode, when autistic children thrive better with color schemes that match their preferences and engage them further.
Solution: Our redesign includes various color palette options to be applied throughout the lesson, such as bright-vivid colors, and muted pastel colors based on the child's preference.
The Results
2 distinct user flows. 2 different means to successful learning.
"I find it hard to directly compare the two flows because they serve different purposes. The settings flow seems more useful for preventing overload, while the re-engagement flow seems more useful after attention is lost."
User Interview with Professor She Yingying She (佘莹莹), Associate Professor and Master's Supervisor in the School of Film at Xiamen University, specializing in Human-Computer Interaction (HCI), Social Robotics, Virtual Reality (VR), and Game Development
"What I like most about the settings flow is that it supports individualization. What I like most about the re-engagement flow is that it gives a gentle way to bring the child back to the task."
Reflection
What I learned from this project:
Working with an existing design system like Khan Academy gives much more structure/direction as far as layout, typography, brand color kit, etc. I appreciate the clarity that it gives as a designer. I learned new tools such as online extensions that can pinpoint the exact font used in designs, as well as Figma extensions that allowed for the motion graphics like the moving butterflies.
What I'd change moving forward:
Ideally, I'd just want to develop the prototype further to work as smoothly as a real Khan Academy lesson. Then, I'd want the prototype to be used by children on the ASD spectrum/their parents, so we have more data to work off in iterations. We are grateful to have gotten in touch with distinguished professors working in this education sector and their input, but I'd like to know how children approach this interface. Perhaps we could compare how long they take to complete a lesson in full with our prototype versus the existing Khan Academy interface.
Related Projects
