1.0
- Are programs made by humans less prone to bias?
- Purposeful bias: Targeted ads on Tiktok and Facebook are direct towards different age demographics
- Purposeful bias: NRA won’t raise awareness about gun dangers and design ads with goal of selling more guns. Topic of contention with extremes in current political system
- Lack of diversity can lead to issues - not intentional bias, but still reflecting flaws in development process
- Ex: not a lot of diversity in most professional software development groups
- In CSP: form co-ed groups to get perspectives and different backgrounds
Actions
- Case of HP face tracking technology not picking up dark-skinned faces
- Likely not intentional, because HP has no reason to deter certain groups from using their products
- Probably not enough diversity in testing pool
- Still harmful
- Could be corrected after identifying issue by advertising testing towards Black/brown customers
1.1
- Crowdsourcing is very important for projects (reduces bias and shares information from different people, allows for distributed computing)
- Ex: Wikipedia, APIs, Github
- Ex: COVID vaccines. Lots of clinical trials had to include people with diverse traits (race, gender, age, etc.)
Actions
- Crowdsourcing at DN CS classes: Github repo featuring key concept executions from every team since sometimes it’s hard to make the connection between example code snippets and project implementation. Seeing different syntaxes and variations could help with this
- School crowdsourcing: Don’t see how to include non-CS students in redesigning attendance website beyond asking for feedback.