Finding Specific Cases
Discovering Problems and Establishing Evidence in Real-World Scenarios
Analyzing real-world AI system cases that students can relate to is highly effective in
developing their critical thinking skills.
Examples include:
– Recruitment AI: Systems that classify resumes and calculate interview ecommendation scores
– Traffic Flow Prediction AI: Systems that optimize signal systems based on traffic volume data
– Medical Diagnosis AI: Systems that provide diagnostic guidance by analyzing patient
pathology information
These real cases reveal ethical issues (bias, lack of transparency, unclear responsibility,
etc.) hidden beneath the surface.
Process of Problem Discovery
– Identifying Bias: Can/Does AI provide a biased answer or answer that can result in
discrimination or disadvantage to some people?
– Lack of Transparency: Can developers explain how and why the AI outputs the way it
does?
– Is private information properly protected using encryption and other safety measures to keep it safe and not disclosed to any parties that the user has not agreed to share with?
– Unclear Responsibility: Who should be held accountable when model errors cause
damage?
Importance of Establishing Evidence
Instead of merely raising issues, it is essential to present evidence supported by objective data and statistics. For example, if it’s discovered that recruitment AI consistently rates female applicants lower:
– What data shows how much bias occurred
– Whether sensitive variables were excluded during model training
These need to be specifically identified to discuss improvement measures.
