AI Explainers

Current Project

As AI becomes increasingly embedded in everyday life, understanding its underlying principles is crucial, yet educational resources for non-experts remain scarce. To address this, we developed three interactive web-based tools designed to demystify core AI concepts — edge detection, confidence thresholds, and sensitivity — through hands-on, engaging experiences.

Each tool was crafted using user-centered and learning sciences principles, allowing users to interact directly with complex AI processes. For example, “The Art of Edge Detection” visually explains how AI recognizes boundaries in images, while the “Confidence Calibration Explorer” and “Toggling Sensitivity” tools introduce users to the concepts of reliability in AI decision-making and algorithmic trade-offs. These tools empower users to adjust parameters, offering real-time feedback that helps deepen their understanding of AI’s strengths, limitations, and ethical implications.

In a study involving 42 adult participants, we found that engagement with these tools increased participants’ familiarity and confidence in discussing AI concepts. A mixed-methods evaluation highlighted interaction patterns like critical thinking, ethical considerations, and real-world application, showcasing how adult learners actively make sense of AI technologies.

This project not only contributes new educational tools to the AI and XAI communities but also provides valuable insights into designing effective, transparent, and learner-centered AI interfaces. The tools are accessible online, supporting a more inclusive AI literacy framework and encouraging ethical, informed AI engagement across diverse user backgrounds.

Publications
Designing Interactive Explainable AI Tools for Algorithmic Literacy and Transparency (DIS 2024)

Other Resources
AI Explainers

Collaborators: Maalvika Bhat

Confidence Calibration Explorer

The Art of Edge Detection

Toggling Sensitivity

Knowledge Net

Current Project

Knowledge Net focuses on teaching learners about knowledge representations and how computers use them to make decisions. On an interactive touch table, learners can create a semantic network to control the appearance and behavior of an animated character. Four learners can create characters and collaborate to create interesting interaction patterns between the characters in the center of the table. Through the iterative design and testing of this exhibit, we are investigating how learning experiences involving embodied interaction and creative making contribute to fostering AI literacy.

Exhibitions
Griffin Museum of Science and Industry, 2024

Publications
Knowledge Net: Fostering Children’s Understanding of Knowledge Representations Through Creative Making and Embodied Interaction in a Museum Exhibit (C&C 2024).

Fostering AI Lit er acy with Embodimen t & Creativity: From Activity Boxes to Museum Exhibi ts (C&C 2023).

Collaborators: Sophie Rollins, Shannon Sauhee Han, Grace Wang, Katherine Hancock, Jasmin Ali-Diaz, Brian Magerko, Jessica Roberts, Yasmine Belghith

Funding
NSF DRL 2214463

DataBites

Current Project

Data Bites is an exhibit to teach young learners about the steps and processes of machine learning and that computers learn from data. Learners approach a table where they can physically construct a dataset to teach a computer how to recognize sandwiches vs. pizzas. They use large tangible sandwich and pizza fixings (e.g., a pizza pie, pepperoni slices, mushrooms) to construct three examples of pizzas and three examples of sandwiches to include in their dataset. The deconstructed fixings allow for creative exploration and testing, such as investigating how including a pizza with unusual toppings or no cheese affects the outcome. The learners then hit a button to train the algorithm and view their results. They can easily iterate on their dataset to improve their sandwich/pizza recognition or test the boundaries of the algorithm. Through the iterative design and testing of this exhibit, we are investigating how learning experiences involving embodied interaction and creative making contribute to fostering AI literacy.

Exhibitions
Griffin Museum of Science and Industry, 2024

Publications
DataBites: An embodied and co-creative museum exhibit to foster children's understanding of supervised machine learning (C&C 2024).

Fostering AI Lit er acy with Embodimen t & Creativity: From Activity Boxes to Museum Exhibi ts (C&C 2023).

Collaborators: Hasti Darabi Pourshiraz, Dev Ambani, Grace Wang, Brian Magerko, Jessica Roberts, Yasmine Belghith

Funding
NSF DRL 2214463

AI Unplugged

Current Project

We are developing a set of “unplugged” (i.e., no computer needed) AI literacy activities that can be used by educators and families interested in incorporating AI literacy into their curriculum or informal learning activities. These activities require only simple craft materials, like pens, paper, and glue. Each activity description is accompanied by a detailed lesson plan that outlines: materials needed (including customizable printable worksheets), key AI literacy concepts that the activity is intended to communicate, key vocabulary, potential connections to other subjects, reflection questions, and AI literacy design considerations that are incorporated. We hope that this suite of activities can enable learners with a variety of different resources to expand their AI literacy, in addition to enabling families and educators to draw connections between museum exhibits, classroom activities, and at-home experiences.

Other Resources
Unplugged activity resources

Collaborators: Kristin Fasiang, Maalvika Bhat, Hasti Darabi Pourshiraz, Lily Ng, Richard Zhang, Lauren Bichelmeir, Grace Wang

Funding
NSF 2343693

AI Unplugged Overview Video

Large Language Madlibs

Intersecting with math education, in this activity, students learn about independent/conditional probability and how LLMs work. Students engage in a MadLibs-like worksheet activity where they use dice and a provided slide deck to explore how LLMs use probability to generate words to complete a sentence. This activity includes exploration of bias in LLM-generated texts and an introductory session on how LLMs are trained.

Monster Mapping

Students learn about unsupervised machine learning and K-means clustering through a paper-based activity in which they cluster different types of colorful monster characters. This activity has intersections with math education objectives of learning how to read and represent data on graphs (Cartesian plane).

DataBites Unplugged

In this activity, students learn about supervised machine learning by using physical pieces to create datasets to train an algorithm to recognize pizzas and sandwiches.

Data Brokers

Learners engage in an interactive game in which they work together to figure out which users to target with ads for different products via an AI algorithm. To aid in their quest, learners use a sample database that mimics the type of data a company would acquire from a third party data broker in order to decide how to allocate their funds for ads. Learners are prompted to reflect on different user identities and whether algorithms can truly “know” us and our preferences.

Introducing AI

What is artificial intelligence (AI)? Where have you used it before? How do you feel about it? How does it work? In this activity students will explore questions like these through an interactive group worksheet/card-deck activity. This activity is intended to serve as a very high-level introduction to AI and can be used to kick off a class, unit, or workshop with other activities involving AI.

Speculating AI Futures

Current Project

We are engaging in speculative design to explore possibilities for what our interactions with creative AI technologies could look like in the future and consider ethical and societal implications. As emerging AI tools like ChatGPT and DALL-E are changing the landscape of AI creativity, we are asking questions about what these technologies could look like decades into the future. Could we interact with these tools in a way that is more consistent with traditional artistic practice? How might we reckon with the environmental impact of large language models? What impacts could these tools have on creatives and communities in the future?

We have created several novel speculative designs exploring AI futures, conducted a speculative design workshop with 18 research participants through which we identified opportunity areas for HCI to inform creative GenAI, and designed a toolkit for fostering interdisciplinary speculation on AI futures.

Publications:

Generative AI Futures: A Speculative Design Exploration (C&C 2023).

Collaborators: Lauren Lin, Jesslyn Im

Onion AI

Onion AI is a generative AI art tool that separates visuals as layers to remix collaboratively, inspired by participatory design and art-making. In 2035, Onion AI is how communities co-create living artifacts like murals in public spaces. Walls and streets are built of digital displays that afford touch, voice and motion-sensing interaction modalities. In one corner, a kid doodles on a layer using the pen tool and generative AI. Simultaneously, a long-time resident tells the story of moving to this city, which becomes the prompt for another AI-generated layer. Best friends select the sky depicted in the mural and jump to record their excitement, resulting in an energetic apricot-pink horizon. Over time, the mural evolves as people interact with it--and becomes a time capsule for the community.

gAIrden

gAIrden imagines generative AI as an agricultural act. This concept creates a parallel present, considering what AI might look like in the face of climate change making energy resources scarce, impacting the computing power available globally.

gAIrden shifts our temporal perceptions of computational speed and encourages sustainable resource management. gAIrdeners “plant” inputs (text-based or art requests)--but AI-generation is not instant. The pace of growth depends on the complexity of the desired graphic and the energy available at that moment. Once the digital plant is fully grown, gAIrdeners can harvest the output—the finished product generated by AI. The nature analogy creates an anticipation period, followed by a delightful (or surprising) reveal. Post-harvest, gAIrdeners receive an environmental cost receipt. Generated pieces are minted with a badge indicating the sustainability of the production, similar to a LEED certification.

Dynamic Data Portraits

In the age of hyper-personalization, people want art that represents them. A couple picks up the Dynamic Data Portrait for their home: an abstract reflection of personal data from their AI chat history and social media. The art changes as shifts are detected across connected platforms. Hang it anywhere--from the dining room to the bedroom, individuals can see themselves from an algorithm’s point of view, bringing visibility to the invisible.

LuminAI (2022)

2016-2022.

LuminAI is an interactive installation in which a human participant and an AI agent can improvise movement together. The AI agent learns from the user and uses dance theory to inform its choice of movement responses. I led the design and development of LuminAI for 6 years and led numerous studies using LuminAI as a tool to investigate how people co-create with AI and each other in the domain of movement improvisation. Most recently, we engaged dance students in focus groups to explore how they communicated with each other using embodied movement to inform the design of LuminAI.

Exhibitions
2017 ACC Creativity and Innovation Festival, Smithsonian Museum of American History, Washington, DC.
CODAME Art + Tech Festival, San Francisco
Children’s Museum of Pittsburgh, 2017

Publications
Exploring Collaborative Movement Improvisation Towards the Design of LuminAI—a Co-Creative AI Dance Partner (CHI 2024)

AI Meets Holographic Pepper’s Ghost: A Co-Creative Public Dance Experience (DIS 2023)

Moviz: A visualization tool for comparing motion capture data clustering algorithms (MOCO 2020)

Visualizing improvisation in luminai, an ai partner for co-creative dance (MOCO 2020)

Learning movement through human-computer co-creative improvisation (MOCO 2019)

Moving in Virtual Space: A Laban-inspired Framework for Procedur al Anim ation (EXAG 2018)

Designing Co-Creative AI for Public Spaces (C&C 2019)

Designing for Socially Interactive Systems (C&C 2017)

Funding
NSF 2123597

Advised by: Brian Magerko
Collaborators: Milka Trajkova, Andrea Knowlton, Mikhail Jacob, Cassandra Naomi, Lucas Liu, Swar Gujrania

Demo Video

LuminAI at The Children's Museum of Pittsburgh

LuminAI at the CODAME Art + Tech Festival

TuneTable

2017-2022

TuneTable is a tangible tabletop experience where users can co-create music using computer science concepts. The experience is designed to encourage interest formation and facilitate learning in CS. I co-led the evaluation team for five years for TuneTable and led the design team for one year to complete the final iteration of the exhibit.

Exhibitions
Griffin Museum of Science and Industry (Chicago, 2017, 2018, 2022)

Publications
Xylocode: A Novel Approach to Fostering Interest in Computer Science via an Embodied Music Simulation (CHI 2024).

Active Prolonged Engagement EXpanded (APEX): A Toolkit for Supporting Evidence-Based Iterative Design Decisions for Collaborative, Embodied Museum Exhibits (CSCW 2022).

Out of Tune: Discord and Learning in a Music Programming Museum Exhibit (IDC 2020).

Trajectories of Physical Engagement and Expression in a Co-Creative Museum Installation (C&C 2019).

Other Resources
Online game version of TuneTable.

Advisor: Brian Magerko
Key Collaborators: Tom McKlin, Astrid Bin, Anna Weisling, Gerard Roma, Anna Xambo, Jiaxi Yang
Funding: NSF DRL 1612644.

Xylocode

The TuneTable project has resulted in several iterations of exhibit designs. The most recent, Xylocode, allows learners to create arrays of flowers by dragging their finger on a touchtable and create bubbles by moving and tapping a tangible frog. When a bubble collides with an array of flowers, a sound plays. Learners can experiment with adding conditionals to iterate through the arrays and manipulating global variables such as the soundscape.

Video presentation of Xylocode for CHI 2024

Xylocode

During the pandemic, we created an online web app version of Xylocode, which you can play with here.

Demo Video of Earlier TuneTable Iterations

Earlier iterations of TuneTable utilized tangible programming languages for music manipulation.

Blockhead Language

In the Blockhead iteration of TuneTable, tangible blocks are connected on the tabletop surface to create and manipulate music. Participants can place sample blocks (top left) to play sound samples, which can be manipulated using other blocks, such as the loop block (top middle), which will cause a sample to repeat over and over again.

AI Literacy Activity Boxes

My dissertation research focused on how to design informal learning experiences to engage families in learning about artificial intelligence (AI). I conducted an extensive literature review of work related to AI education and synthesized findings from the literature into a set of AI literacy competencies and design considerations. I used these to inform a design-based research project in which I created three activity boxes for informal family learning about AI. These boxed activities used tangible interaction, collaboration, and creativity to engage families in learning about a variety of AI concepts ranging from supervised machine learning to understanding knowledge representations. The activity box prototypes were summatively evaluated via at-home user testing due to COVID-19.

Publications

Conducting Remote Design Research on Embodied, Collaborative Museum Exhibits (CHI 2023)

Family Learning Talk in AI Literacy Learning Activities (CHI 2022).

Co-Designing AI Literacy Exhibits for Informal Learning Spaces (CSCW 2021).

The Role of Collaboration, Creativity, and Embodiment in AI Learning Experiences (C&C 2021).

What is AI Literacy? Competencies and Design Considerations (CHI 2020).

Other Resources

Exhibit designs and resources

Advised by: Brian Magerko

Collaborators: Aadarsh Padiyath, Anthony Teachey, Jonathan Moon

AI Literacy Activity Boxes

I designed three activity boxes to engage families in learning about AI: 1) Knowledge Net, which teaches about knowledge representations by engaging families in creating a semantic network to teach a chatbot about the world, 2) Creature Features, which engages learners in creating a dataset for bird recognition algorithm, and 3) LuminAI, an AI dance partner with an interpretable interfaces and visualizations to explain the AI’s reasoning processes.

Stage 1 - Defining AI Literacy

I conducted an extensive literature review of work related to AI education and synthesized findings from the literature into a set of AI literacy competencies and design considerations

Stage 2 - Low-Fidelity Prototyping

I developed a design workbook with ideas for 25 different AI literacy museum exhibits. This design workbook was inspired and organized according to the AI literacy competencies and design considerations developed in Stage 1. I developed low-fidelity paper-based prototypes of 5 of these ideas.

Stage 3 - Co-Design

I conducted a co-design workshop with family groups at the Griffin Museum of Science and Industry. Family groups engaged in worksheet and design activities to share their experiences and feelings about AI and brainstorm ideas for AI literacy museum exhibits. They also interacted with and provided feedback on several of the low-fidelity prototypes I developed in Stage 2.

Stage 4 - Iterative Design Research

I iteratively developed and tested three different AI literacy activities for informal learning spaces—1) Knowledge Net; 2) Creature Features; and 3) LuminAI. Due to COVID-19, I developed these as box-sized exhibits that could be delivered to families homes.

User Testing

I tested these exhibits by boxing them up and delivering them to families around the Atlanta metro area. I conducted remote user studies with families to assess their learning, interest development, and level of collaboration/creativity/embodiment in their interactions.

Bee a Pollinator!

2019

Bee a Pollinator! is an exhibit designed to teach kids about pollination via a playful, interactive experience. Families are invited to use stuffed felt bee wands to pick up colorful velcro “pollen” balls. If the player moves the pollen ball to the matching colored flower hole, the stem of the standing flower begins to light up. After three balls have been added, a fruit in the center of the flower moves up and down. This experience was designed to communicate three key learning goals about pollination: 1) bees pollinate flowers; 2) pollen sticks to bees; 3) flowers are only pollinated when a bee leaves behind the “matching” pollen; and 4) pollinated flowers produce fruit. The installation uses USB cameras and a color recognition program to detect the color of the pollen balls and an Arduino with NeoPixels and servo attachments to power the lights and fruit movement.

Exhibitions
Children’s Museum of Atlanta, 2019

Exhibit chosen to be a permanent installation at the Children’s Museum of Atlanta (postponed due to COVID-19)

Advisor: Hyunjoo Oh
Collaborators: Himani Deshpande

Players move colorful pollen balls to matching flower holes in order to make the stems of the flowers light up. After three balls are added, a fruit pops up from the center of the flower.

We custom-sewed stuffed bees out of felt. When participants stuck one of the bees in the hole in the installation, velcro “pollen” balls would stick to it.

The installation structure was constructed from a combination of plywood and cardboard.

Flower petals

The petals of the standing flowers were constructed by wrapping custom made aluminum and steel wire frames with colorful stocking material.

Fruit Mechanism

We used a rack and pinion mechanism (laser cut out of acrylic using designs from Paper Mech) attached to a servo motor to move the plastic fruit in the center of the flower up and down. The servo was controlled by an Arduino board.

In order to light up the stems while still maintaining the appearance of a flower, we wrapped NeoPixel LED lights around skinny PVC pipes, which we then inserted into a larger, clear PVC pipe wrapped in green floral tape. The result was a semi-translucent green stem with lights inside that matched the color of the flower petals.

Personal Equity Index

2018

The City of Atlanta is exploring how to represent and address issues of equity in the city, and the Personal Equity Index is a participatory workshop designed to aid in humanizing statistics and quantitative data with qualitative stories about (in)equity. The workshop, developed through an iterative design research process in collaboration with the City of Atlanta, is centered around the idea that equity is personal. In the workshop, participants use workbooks and card decks that we designed to explore how they feel about issues of equity and inequity in Atlanta. The insights from this workshop were intended to enable the City of Atlanta to examine the qualitative stories and personal experiences of residents in Atlanta on both on an individual and community wide scale.

We structured the workshop around three key ideas: intersectional identity, personal experiences, and emotions/feelings. Participants are first asked to share three words that best describe who they are, as a way of expressing their own personal, intersectional identity. They are also asked to situate themselves in a neighborhood that they live in Atlanta. They are then asked to discuss their experiences living in Atlanta, focusing both on things they are grateful for and things they wish they could change about the city. Finally, they are asked to discuss their overall feelings about living in Atlanta, and share a story of a time the city made them feel a certain way (e.g. included, vulnerable).

Advisor: Carl DiSalvo
Collaborators: Sucheta Ghoshal, Bang Tran, Xu Zeng, City of Atlanta

Workbook

The workbook prompts participants to use the cards to explore their personal stories of equity in Atlanta. Participants use the Experiences cards to explore what they are grateful for and what they wish they could change in Atlanta, and they use the Feelings cards to explore a time they have felt a certain way (e.g. vulnerable, included) in Atlanta.

Feelings and Experiences Card Decks

The Feelings card deck consists of emotion words like vulnerable or included. The Experiences card deck is broken down into six sub-decks, each of which focuses on a different aspect of the City of Atlanta’s agenda - Transportation, Safety, Health and Wellbeing, Empowerment, Inclusivity, and Housing.

City of Atlanta Workshop

We held a pilot workshop with a group of City of Atlanta officials, who participated in a one-hour workshop and offered feedback on our work.

Sound Happening

2017-2018

Sound Happening is an installation in which participants can collaboratively create music by moving colorful balls around an interaction space. A webcam tracks the location and color of the balls and generates music according to those parameters. My role on this project focused on iterative installation design and evaluation.

Exhibitions
The Children’s Museum of Pittsburgh, 2017
2019 ACC Creativity and Innovation Festival at the Smithsonian Museum of American History.

Publications
Don't Steal My Balloons: Designing for Musical Adult-Child Ludic Engagement (IDC 2018).

Advised by: Brian Magerko
Key Collaborators: Richard Savery, Raianna Brown, Dan Singer, Pedro Arevalo, Josh Moore, Agneya Kerure

Sound Happening at the Clough Art Crawl, Georgia Tech

Performance at 2020 ACC Creativity and Innovation Festival, Washington D.C.

Interactive Installation

Diagram of Installation Design

The Shape of Story

2017

The Shape of Story is an interactive story circle experience in which participants collectively create a story line-by-line. In the installation, fabric panels delineated a circular room where participants sit in a circle on pillows surrounding a circular projection screen. Participants speak one by one into a sculptural listening device, each telling a line of a story. As participants collectively tell a story, a visualization is projected in the center of the circle. Artificial intelligence in narrative understanding is used in conjunction with a symbolic visual language to visualize the story in real-time. Every element of the visualization, from color to shape to the speed at which lines are drawn, corresponds to an element of the story that is told. The result is a communally created narrative art piece.

Exhibitions
Eyedrum Art and Music Gallery (Atlanta, 2017)

Publications
The Shape of Story: Semiotic Artistic Visualization of a Communal Storytelling Experience (INT Workshop at AIIDE 2017).

Code
Github

Advised by: Brian Magerko
Collaborators: Sanjana Gupta, Jessica Brooke Anderson

Story Visualization

Storytelling Experience

Symbolic Language

Every element of the visualization, from color to shape to the speed at which lines are drawn, corresponds to an element of the story that is told.

Physical Installation Design

Fabric panels (right) delineate a circular room (center) where participants sit in a circle on pillows (left top) surrounding a circular projection screen (left bottom). As participants collectively tell a story, the visualization is projected in the center of the circle.

Speaking Device

With the help of local Atlanta artist Jessica Brooke Anderson, we constructed a sculptural speaking device. This device is passed around the circle, and participants use it to share their piece of the story.