Does AI recognise the diversity of our world
Using the example of image-generating AI, learners explore participation and inclusion in the digital sphere. They compare an AI-generated image of Vienna’s population—or that of another city, using appropriate local data—with statistical information and examine which social groups are represented, underrepresented or absent. In doing so, they reflect on how biases in AI training data can shape digital representations of society and develop their ability to critically evaluate AI-generated content.
| Age group | 11 – 14 years |
|---|---|
| Duration | approx. 50 minutes |
| Learning goals & outcomes | Learners compare an AI-generated image with statistical data and use this comparison to reflect on how different groups are represented in AI-generated images. In this way, they develop awareness of participation and inclusion in the digital sphere. |
| Methods | Analysis of an image and of charts presenting statistical data |
| Source / Author | Lorenz Prager, Digital Citizenship Education: Grundlagen und Lehr-Lern-Settings zur digitalen Mündigkeit (Vienna, 2025), available at: https://www.politik-lernen.at/DigitaleMuendigkeit. |
| Values | Equality: Learners examine whether different social groups are equally represented in AI-generated images and reflect on how biases can lead to exclusion and unequal visibility in the digital sphere. |
Didactic notes
This teaching and learning setting addresses inclusion within the digital sphere. By developing factual and value judgements, it aims to raise awareness for participation and inclusion. As generative AI currently attracts considerable attention, the topic provides a useful starting point for exploring inclusion while also strengthening learners’ media literacy in relation to AI. The lesson uses data from Vienna but can be adapted to other cities and municipalities.
Material & Downloads
Information sheet; worksheet; an account at an image-generating AI provider or the provided AI-generated image
Introduction and guiding question
Begin by demonstrating an image-generating AI system. Using an AI tool of your choice (e.g. Midjourney, Stable Diffusion or DALL-E via ChatGPT), enter the prompt: ‘Generate an image showing 50 people who live in Vienna.’ A ready-made image is provided in the materials (Worksheet M1, ‘AI-generated image’) in case of technical difficulties. Then explain that the aim of the lesson is to assess whether the image represents everyone. If learners ask whether this is relevant, use the opportunity to explain the importance of inclusion in both analogue and digital spheres. In brief, mass media play an increasingly important role in how people perceive their social environment. They shape our view of the world, which is why the representation of different groups in the media matters in an inclusive society.
Background information and distribution of worksheets
Next, explain how image-generating AI works. The level of detail can be adapted to the learners’ age (see the information sheet ‘Image Generation’). The essential point is that AI systems use training data and can reproduce biases contained in those data. Then distribute the worksheets (‘Tasks’, M1 ‘AI-generated image’ and M2 ‘Statistical material’) and discuss the steps required to complete the activities.
Individual work and discussion
Learners first develop factual and value judgements individually and then discuss them with the whole class. For citizenship education, it is important that learners do not merely form value judgements and stop at their individual positions. These judgements should also be communicated, compared, examined and discussed. This process encourages learners to reflect on how they form their own value judgements.
Variation
In the form presented here, the teaching and learning setting is designed for learners aged 13–14, but it can also be used with older age groups. For differentiation within or between groups, the tasks can be modified to provide additional support or require more independent work. For example, learners can generate the images themselves using AI, or they can be given access only to the statistical yearbook and asked to conduct their own research.
Lesson material
The lesson materials are provided in a simple Word format so that teachers can adapt and modify them.