1 of 18

Teaching Adolescents about the Privacy Threats of Tracking and Pervasive Personalization:

A Classroom Intervention Using Design-Based Research

Sushmita Khan

Humans and Technology Lab (HATLab)

Clemson University

2 of 18

The HATLab

Situated in the Human-Centered Computing program in Clemson’s School of Computing, the HATLab is a multidisciplinary space where students, faculty, and researchers come together to investigate all aspects of humans’ interactions with technology.

Topics we research:

human-centered computing
human-computer interaction
AI and adaptive systems
usable privacy and security
health informatics
privacy-enhancing technologies
designing for special populations
cross-cultural research
equity and inclusion
innovation in empirical methods

Cryptographic Provenance for Digital Publishing

(NSF SaTC Award 1940679)

Mathematically-grounded metaphors to teach AI-related cybersecurity

(NSF SaTC EDU Award 2039616)

3 of 18

The CyberEd Project

A four-year collaborative project between College of Education and School of Computing that developed and assessed learning modules to teach middle school students AI cybersecurity principles and best practices within the context of their existing math and computer science courses.

Developed a design-based approach to teaching cybersecurity

Sushmita Khan…Emily Sidnam-Mauch, Nicole Bannister, Kelly Caine, and Bart Knijnenburg (2024). Teaching Middle Schoolers about the Privacy Threats of Tracking and Pervasive Personalization: A Classroom Intervention Using Design-Based Research. ACM CHI '24.

Sushmita is building upon this work in a new context for her dissertation

Scan for Paper

4 of 18

Objective

Overview Background and Case Study:

Applying design-based research to teach middle school students about cybersecurity threats.

Discuss Next Step:

How can we build AI and privacy awareness into data collection sites (whether an organization, website, etc)?

5 of 18

Classroom Research

Creating cybersecurity learning modules in a classroom is useful because it…

Broadens access
Promotes collective adoption of privacy-preserving practices
Sustainable knowledge development

6 of 18

Classroom Research

Deploying modules in a classroom is hard because it is an…

Uncontrolled environment with many moving parts

7 of 18

Design-Based Research (DBR)

On conducting classroom interventions

8 of 18

DBR Approach

9 of 18

We identified 5 learning needs for middle school students:

AI Perception
AI Recognition
How AI works
AI Engagement
Privacy Concerns

10 of 18

Education Modules

AI Agents
Tracking and Pervasive Personalization
Virality
Misinformation
Filter Bubble
AI Bias

11 of 18

Learning Activities

Interact with personalized data dashboard
Design a targeted ad

12 of 18

Learning from the education intervention

Takeaways:

Rationalize Algorithmic Decision
Association between data, algorithmic decisions and privacy
Cautious and Deliberate Interaction

13 of 18

Next steps..

What I’ve learned from DBR

Conversation of Privacy in context is more productive

Where I want to take it with my dissertation

Contextualizing Privacy in Personal data

the messiness of real world practice must be recognized, understood, and integrated as part of theoretical claims if the claims are to have real-world explanatory value.

Context is an integral part of the complex causal mechanisms that give rise to the phenomenon under study

Students learn best when it is contextually relevant and take it to the next level and connect it to their own data

If researchers only study that which takes place in controlled laboratory conditions, they run the risk of developing artificial meanings and interactive dynamics so free of contextual realities that any interpretations are suspect with respect to informing real-world practice. Because learning scientists are committed to producing change in real-world learning environments, learning sciences research often occurs in naturalistic contexts and includes rich descriptions of the conditions through which the research was conducted

14 of 18

Contextualizing Privacy in Personal Data

Inferences

Recommendations

15 of 18

How can we build cybersecurity and privacy awareness into data collection sites (whether an organization, website, etc)?

Discussion

16 of 18

Discussion…

What do you see as the greatest AI-related cybersecurity and privacy threats, and how would you prepare young adults to be aware of these issues and in charge of their online identities?
How can organizations and data collection systems be designed to help individuals be more aware of these cybersecurity issues and how they impact individuals?
Are there practices you use in your company, features you’ve embedded in a product, or research you’ve conducted that make individuals aware of threats to their privacy or information security?

17 of 18

Thank you!

Sushmita Khan, sushmik@clemson.edu (HCC Ph.D. Candidate, Clemson University)
Emily Sidnam-Mauch, esidnam@clemson.edu (Research Scientist for the Clemson HATLab)
Learn more about HATLab at https://hatlab.org/

18 of 18

Findings

Learning needs
Education Modules
Learning from the education intervention