<aside> 🍵

note that this part of the unit has been recently added to the curriculum. we predict it’ll show up on a few MCQs so we’ll summarize key points here! hopefully this is pretty much self-explanatory~

</aside>

★ data privacy

phones, computers, and other devices and software have lots of our personal data, which we don’t realize is a risk to our personal privacy. programmers need to be careful and uphold users’ personal privacy, especially when their code is involved in collecting and storing data in computer systems…

• legally and ethically, you must ask the users for permission to access and store their data
• if a breach happens (where data is stolen), you must inform the affected users

there are typically pros and cons when programs have access to our data, security wise

• for example, software for maps use our location. it can be helpful if you’re lost and need to be found, but it can also be unsafe if someone unauthorized gains access to your location

(✿◠ᴗ◠) make sure you can identify other examples of data collecting and privacy issues

★ data and bias

artificial intelligence (AI) and machine learning (ML) rely on huge datasets and lots of energy, which raises ethical issues about where this data comes from, privacy, and resource use

dataset*:** collection of specific pieces of information or data*

• particularly, internet data contains societal bias so AI trained on it can learn and hence reproduce those biases (i.e. CEO portrayed as white male)
  • this can be harmful and create unfair outcomes in real life
• algorithmic bias: systematic, repeated errors that disadvantage a particular group
  • automated hiring tool trained on employee data becomes biased against female applicants
  • facial recognition being less accurate for darker skin tones because of underrepresented training data
• in order to reduce bias, programmers should…
  • use diverse/representative datasets; include diverse teams in development and testing
  • test for bias and accuracy regularly
  • be transparent about limitations
  • follow ethical guidelines and standards

(✿◠ᴗ◠) make sure you can explain why data quality and bias matter; find good/bad datasets and how it can affect results