|

Process Post 3: Bias and Fairness in AI for Education

felipemottaf

Learning Focus

This week I studied how bias appears in AI tools used in education. AI systems learn from past data, and that data can reflect unfair patterns. My goal was to understand how this affects students and what teachers can do to reduce harm.

Learning Activities and Tools Used

I read a UNESCO section on AI ethics that explained how biased data can shape student feedback, grades, or learning paths. It said that AI must be checked often to make sure it treats all students the same.

I also looked at a paper by Buolamwini and Gebru (2018). They tested AI models that detect faces and found that the tools worked better on lighter skin tones than darker ones. This showed me how bias in training data can lead to real problems when AI is used for learning, monitoring, or assessment.

I also read a study by Holmes et al. (2022) about the need for human oversight. They explained that teachers must stay involved in AI-supported decisions to prevent errors or unfair outcomes.

Reflection

This topic made me rethink how “neutral” AI really is. AI can look objective, but it still carries the choices and limits of its designers and datasets. If an AI tool gives poor feedback to certain groups of students, it can affect their confidence and progress.

I also learned that fairness is not automatic. It must be built into the system. Teachers need simple tools that show how AI makes decisions. Students should also know if an AI tool might misread their work or behaviour.

Next, I want to study how schools can check AI tools for bias before using them. I also want to learn what rules or standards help make AI systems safer and more fair.

References