UNIT17

Transcript

1. culture defines reality. and language determines how we interpret it. culture defines reality. and language determines how we interpret it.
2. our language is androcentric Mr. can refer to any man, regardless or whether he is single or married, but Miss and Mrs. define women by their civil status. Male is generic: mankind, manmade. "you guys"
3. why is this a problem? The language we use not only reflects our culture, but also constructs it. words matter.
4. try to avoid gendered pronouns how can we solve it? fireman - firefighter policeman - police officer postman - post officer freshman - first year student
5. use more inclusive forms.
7. WORDS ARE TOOLS OF THOUGHT. WE CAN USE WORDS TO MAINTAIN THE STATUS-QUO OR TO THINK IN NEW WAYS .
8. until now we have been forced to speak the language of computers. now computers are learning how to speak ours. MACHINE LANGUAGE
9. what are we going to teach AI? what are we going to "input"? and on what rules is the output based? MACHINE LANGUAGE
10. Should AI be gendered? Have human values? Should it be religious?
11. father : doctor :: mother :nurse man : computer programmer :: woman : homemaker WORD EMBEDDINGS a popular framework to represent text data as vectors, used in many ML and NLP tasks.
12. RECRUITING Tools such as Applied, GapJumpers, and Unitive allow employers to blind themselves to applicants’ demographic characteristics & strip age, gender, educational and socioeconomic background, so they can focus on talent only. #nurturing #supportive #assertive #competitive
13. SELF DRIVING CARS Seat belts were designed to be safe for men and, for years, we sold cars that were largely unsafe for women, especially pregnant women Consequently, female drivers are 47% more likely to be seriously injured in a car crash. In 2011, the first female crash test dummies were required in safety testing — but we are still building on 50+ years of dangerous design practices for automobiles.
14. AMAZON ECHO LOOK You cool with an algorithm, machine learning, and “fashion specialists" deciding whether you look attractive today? What sorts of built-in biases will an AI fashionista have? It's worth remembering that a recent AI-judged beauty contest picked primarily white winners.
15. AMAZON ECHO LOOK You cool with an algorithm, machine learning, and “fashion specialists" deciding whether you look attractive today? What sorts of built-in biases will an AI fashionista have? It's worth remembering that a recent AI-judged beauty contest picked primarily white winners.
16. how can AI support people in making less biased decisions?
17. ASKING WHY MAKING RECOMMENDATIONS CORRECTING FOR SLANG/ NON-STANDARD DIALECTS ADDING MORE RULES TO THE ALGORITHM
18. thank you ! @andra_bria