Lesson 4: Bias, fairness and privacy in the inbox
61 min read·14 Oct 2026
Module 10 · Lesson 4 · About 60 minutes · Needs: basic Python; the TF-IDF classifier from Module 4 Lesson 4 and word vectors from Module 6 help, but this lesson rebuilds what it uses
The situation
It is Wednesday 11 November 2026, project week 10. Rosa wants the TF-IDF router live. Amara asks two questions first: does it treat every customer the same, whatever their name or language, and what personal data does it keep?
The running example is HT11 (simulated): a long, polite message about battery batch VB-0426, signed "Many thanks, Margaret Ellis". Ines labels it warranty_fault, for Dev. Change only the name and watch where it goes.