An unplugged educational activity in which students train a simulated AI-based recruitment filter using training data, identify patterns, apply these patterns to new applications, evaluate the resulting decisions and critically discuss whether the criteria used by the AI are appropriate.
Why it matters for ENACT-AI
The activity introduces pattern recognition, training data, automated decision-making and potential bias in AI systems.
Learning outcomes
Understand how AI systems can identify patterns in training data, understand the basic logic of training an AI-based decision system, recognise how training data influence automated decisions, evaluate AI-generated decisions critically, identify potentially problematic or discriminatory criteria, discuss fairness and responsibility in automated decision-making.
Suggested activity
AI Recruitment Filter Challenge, students work with the provided fictional recruitment data and gradually train the simulated AI filter. They identify patterns, compare datasets, adjust the model, apply it to new candidates and then critically evaluate whether the resulting decisions are fair and appropriate. As a final step, students define criteria for what a “good” AI recruitment filter should look like.