Understanding Bias in Algorithms

Grade 9s looked at several articles today and we practiced marking up text and summarizing. We talked about strategies good readers use to interact with the text, and then we broke off into smaller groups jigsaw style to read and learn from several articles.

One article talked about pulse oximeters registering different values depending on the skin’s melanin content (because melanin is interpreted as higher hemoglobin counts and higher oxygen content of blood), so algorithms and devices have a built in bias. Another article talked about how AI can be a big help, but if it is trained on biased materials, it will reinforce and amplify the bias. Another talked about how data security is important, and how people should consider what health information is entered onto apps, using the example of period tracking apps in the United States where abortion is becoming more illegal in certain states. Another talked about data security of apps like tiktok where foreign countries can have influence and access.

We had some interesting discussions, and some students are rethinking which apps they are using, and who may have access to their information. We also recognize the importance of having a diverse group of people represented at the coding table!