IB Digital Society 5.2C Diversity and Discrimination Notes

5.2C Diversity and Discrimination

Discrimination as a Global Challenge

1. HL-only extension: how digital systems interact with diversity and discrimination across gender and race, disability and inclusion, and tolerance or hostility online.
2. Discrimination is older than any technology. Digital systems rarely invent prejudice; they change its reach and speed, scaling it up and hiding it inside a model that looks neutral.
3. Anchoring concepts: values and ethics and power (who a system treats fairly and who decides), plus identity, since systems act on the traits discrimination targets (race, gender, disability).

Bias Built In

1. Algorithmic bias: a system deciding who gets an interview, a loan, or extra police attention can treat some groups worse, even when no one intended it to.
2. Bias usually starts in the data: a model learning from unfair records repeats the pattern, and because a computer produced it, the unfairness can look objective. Cases: hiring tools that downgraded women's CVs; facial recognition that misidentifies darker skin far more often.
   1. Through power, the label a system assigns quietly decides a person's chances; through values and ethics, a machine-made decision can still be deeply unjust.
   2. Impacts vs implications: real people are denied jobs, loans, or fair treatment, and the danger is discrimination scaled up and hidden. There is one opening, though: once bias is found in an algorithm, it can be measured and corrected in ways a human decision-maker's bias cannot.

Being Seen and Served

1. Discrimination is also who a system sees and serves: recognition tools work well for some groups and poorly for others, treating people unequally at the point of contact.
2. Accessibility features open digital life to people otherwise shut out (screen readers, captions, alternative controls); when missing, the same systems build new walls.