Contents
- 1 Goals
- 2 Steps
- 3 Variations
- 4 Reflection
Goals
Participants learn:
- how AI and algorithms use personal data to make predictions
- the difference between self-expression and machine interpretation
- the risks of profiling and data-based assumptions
- the importance of privacy, context, and nuance in digital identity
- how to think critically about the “data stories” that shape our futures
Personal Data and Predicted Identity
In a world increasingly shaped by algorithms, our personal data is used not only to reflect who we are — but also to predict who we might become. This method invites adolescents to explore how digital systems interpret identity by experimenting with AI-generated "future self" portraits. Starting from personal or fictional self-descriptions, participants are introduced to how predictive models work: how they gather patterns from data, make assumptions, and often simplify or distort human complexity. Through comparison between their own imagined futures and those constructed by an AI, young people reflect on the difference between self-expression and algorithmic profiling.
By blending creativity, critical thinking, and data literacy, this activity fosters a deeper understanding of how identities are shaped — not just by ourselves and others, but increasingly by unseen digital systems. It challenges participants to consider: What does my data say about me? What does it miss? And how can I take back agency in a predictive world?
This activity is designed to help young people explore how data can be used to predict, shape, or even misrepresent identity. It opens space for creative and critical reflection around the following questions:
- What story does my data tell?
- What can machines see — and what do they miss?
- How do algorithms interpret who I am and who I might become?
- How accurate or biased can these predictions be?
Steps
What We Did
We started with self-descriptions as teenagers — how they see themselves, what they share online, and what they choose to hide. From these personal snapshots, we asked an AI agent (ChatGPT, GPT-4) to generate a predictive portrait of what their lives might look like at age 35.
The AI created a future version of each character based on patterns in their current interests, behaviours, and contexts — for example, Paul’s passion for gaming and photography, or Anna’s data-driven lifestyle and athletic discipline. The AI also explained the process behind these predictions, using simulated data clustering, behavioural inferences, and correlation with similar digital profiles.
- See example profiles of Paul and Anna
1. Warm-up Discussion (10–15 min)
Prompt questions:
- How do you present yourself online?
- What do you choose to show — and what do you hide?
- Have you ever felt misunderstood because of how you look or post online?
- What do platforms really “see” about us?
Introduce the idea that identity is constructed not only through human interaction, but also through algorithms, data trails, and predictive profiling.
2. Activity: My Digital Self-Portrait (20–30 min)
Participants create a digital self-portrait, representing:
- What they usually show online (avatar, stories, selfies, emojis, playlists…)
- What they hide or protect — and why
- How others may perceive them
- How they believe an algorithm might describe or rate them
Working Sheet
Tip: Use a guided worksheet with zones like:
- “Visible to others”
- “Hidden from others”
- “How I want to be seen”
- “What the algorithm might assume”
3. Mini Input: What Do Algorithms See? (10–15 min)
Facilitator explains (briefly) how platforms track behaviour and create profiles. Use real-world examples: recommender systems, biometric tracking, credit or insurance scoring, etc.
- What is a “predicted self”?
- What is the “unknown self” in data?
- How are recommendations, ads, or scoring systems built from our data?
For background knowledge and inspiration, consult the handbook: More than going with the flow - chapter 2 Identity, and example profiles
Ask:
- What do algorithms get right?
- What might they get completely wrong?
4. Activity: My Future Self vs. AI-Predicted Self (30–40 min)
4.1 Personal Future Vision
Participants imagine themselves at 35:
- What kind of life do I want?
- Where do I live, work, spend my time?
- What values do I want to live by?
- What matters most to me?
4.2: The Predicted Self
They receive (or build) an “AI-generated” profile based on limited data — inspired by Paul/Anna-like cases. Ask participants to compare:
- Is this the future I want?
- What does the AI assume? Why?
- What is missing from the AI's version of me?
- How do we resist being reduced to a profile?
Optional Extensions
- Data pen-pal activity (inspired by Dear Data): exchange data-based portraits with a peer
- Create a visual wall: “Real Me / Predicted Me / Hidden Me”
- Focus on platform design: What encourages self-expression vs. conformity?
- Link to digital rights: Explore topics like surveillance, algorithmic fairness, or data justice
- Use AI tools (like ChatGPT) to simulate “algorithmic profiling” — then critique it!
Variations
This activity can be developed in several different ways, depending on your goals and group:
Self-Portrait & Privacy Awareness
Participants create self-portraits based on what they show online vs. what they hide. They reflect on how their privacy choices influence how they are perceived or profiled.
Imaginary Profiles
Participants invent fictional digital identities and imagine how an algorithm might interpret them. This highlights how limited data can lead to oversimplified or biased predictions.
Future Self vs. Predicted Self
A classic “future self” exercise — often used in career or life planning — is contrasted with an AI-generated prediction. This comparison sparks discussion about personal agency vs. algorithmic assumptions.
Reflection
Plenary discussion:
- What surprised you in this activity?
- Have you ever been judged based on what you post or don’t post?
- How do platforms encourage or discourage certain versions of ourselves?
- How do we stay in control of who we are online — and offline?
Close with the question:
- Who gets to decide who I am — me, others, or the system?
Learning Outcomes
By the end of the workshop, participants will:
- Reflect on how they construct and control their digital identity
- Understand basic concepts of data profiling, predictive algorithms, and algorithmic bias
- Explore the impact of datafication and surveillance on self-image and freedom
- Think critically about future possibilities — not just what is predicted, but what is desired
- Build awareness of the difference between tracking for self-awareness and pressure to perform
