AI & Us Exploring Global Challenges through the ‘eye’ of AI’. 

On Tuesday, 27th of October we ran a practical session, AI & Us Exploring Global Challenges through the ‘eye’ of AI’, for academic staff and students at the University of Aberdeen as part of their 'AI Learning Labs' initiative, which provides workshops to explore using Generative AI, with real-world examples.

This hands-on, thought‑provoking workshop invited participants to blend creativity with critical reflection as they explored how generative AI can shape our understanding of global challenges.

Activity Design

The learning activity we used was based on ‘AI Explorers: Investigating the UN Sustainable Development Goals’ which forms part of our projects Generative Artificial Intelligence Skills in Schools (GenAISiS) and Generative Artificial Intelligence Skills in Schools: Teaching, Reaching and Activating – Instructors Network (GenAISiS T.R.A.I.N.), funded by Responsible Ai UK. The activity aims to help learners explore global challenges while critically engaging with generative AI tools. Participants begin by selecting one of the United Nations Sustainable Development Goals (SDGs), such as ending poverty or achieving gender equality. 

A central feature of the activity is the use of the ‘Six Thinking Hats’ framework (De Bono, E., 1985, ‘Six Thinking Hats: An Essential Approach to Business Management’. Little, Brown, & Company), which structures critical discussion. 

Each “hat” in the text based activity represents a different perspective:

• the white hat focuses on information and prompts used,
• the red hat explores transparency and representation,
• the yellow hat considers bias,
• the green hat addresses privacy, trust, or data safety,
• the blue hat reflects on usefulness or emotional response, and
• the black hat supports critical evaluation and overall judgement.

Working in teams, learners use these perspectives to analyse both the AI-generated outputs and the process of using the tool. They are encouraged to question accuracy, representation, fairness, trustworthiness and potential risks.

Finally, participants record their ideas using visual ‘sticky notes’, supporting collaborative reflection and discussion. 

The activity combines AI literacy, critical thinking, and global citizenship education, helping learners understand not only how AI works, but also how it can be used responsibly when engaging with complex real-world issues.

Activity Outcomes

Participants worked in teams to generate and analyse AI-created images linked to one of the United Nations Sustainable Development Goals (SDGs), with a preference on environmental issues and gender equality. Instead of using 'sticky notes' we used Padlet to allow for more interactivity and collaboration.

Using structured prompts and guided discussion, the session unpacked how AI represents people and places, where bias emerges, and how trust is built or undermined. 

🤖 AI Hallucinations

It was interesting that see that several of the AI generated images, two years after running this activity with children, AI is still producing hallucinations. These examples below offer instances of where the AI produced visual inaccuracies or impossible details. For example, participants noted that:

• Aberdeen skyline: it is not the real cityscape, with repeated houses.

• Office scene: workers appeared both inside and outside the glass walls.

• Marine life: whales were in shallow coral reefs, turtles with distorted anatomy, fish with two tails or no heads.

• Human figures: one person was shown with three legs!

AI hallucinations are errors in realism or logic AI may produce. AI can invent things that do not exist or can misrepresent physical reality.

🎯 AI Bias

Bias was also highlighted by participants. Bias means that the training data of the AI defaults to reflecting limited perspectives that can be idealised or geographically narrow. AI bias can reinforce stereotypes or narrow viewpoints. For example, in the AI outputs several instances of bias were identified:

• Over‑representation of tropical or idyllic ecosystems when illustrating global goals.

• Preference for certain species (like turtles) or Western imagery.

• Lack of pollution or hardship, showing only “thriving” environments.

• Generic or marketing visuals that ignore local context or diversity.

Through lively conversation and visual reflection, participants sharpened their AI literacy, while gaining insight into responsible and ethical engagement with AI technologies.

The workshop was part of the project Generative Artificial Intelligence Skills in Schools (GenAISiS) funded by Responsible AI UK. A full set of this and other activities can be found in GenAISiS Instructor Guide.

Thank you to the AI Learning Labs at University of Aberdeen for the kind invitation!