Data Ethics for Regular Developers - rubyconf 2018

Transcript

1. DATA ETHICS FOR REGULAR DEVELOPERS: WHAT WE CAN LEARN ABOUT DATA ETHICS BY PRETENDING TO NOT KNOW WHAT PRINGLES ARE ONLINE COLIN FLEMING @RIDETHECATBUS
2. CONTENT WARNING
 • DISCUSSION OF DEATH • DISCUSSION OF ABORTION CARE
3. IMAGINE AN INVENTION THAT MAKES TELEPORTATION POSSIBLE
4. BUT THERE IS A COST
5. SURPRISE! THE INVENTION IS CARS
6. DOUBLE SURPRISE! PEOPLE’S DATA IS FUEL
7. HMM WHAT’S IN THE NEWS TODAY OH DEAR:
 • EXPERIAN BREACH? • FACEBOOK/GOOGLE BREACH OF THE MONTH?? • PALANTIR ROBOCOP ALGORITHMS???
8. REGULAR DEVELOPERS STARE DOWN THESE ETHICAL QUESTIONS ALL THE TIME
9. OUR ROADMAP FOR THIS TALK:
 • CRASH COURSE IN ETHICAL REASONING • IDENTIFYING OUR VALUES • A FEW CASE STUDIES
10. ETHICS IS OUR RESPONSIBILITY TO ACT IN A WAY INFORMED BY SHARED AND PERSONAL VALUES (MORE OR LESS)
12. PROGRAMMER CODES OF ETHICS (LIKE THE ASSOCIATION FOR COMPUTING MACHINERY’S) ARE A THING!
13. ACM ETHICS SUMMARY: • EVERYONE IS A STAKEHOLDER IN COMPUTING • AVOID HARM AND PROTECT PRIVACY • AVOID YOLO-DRIVEN DEVELOPMENT
14. THE DC ABORTION FUND’S VALUES (ROUGHLY):
 • EMPOWER CLIENTS TO CARRY OUT THEIR CHOICES • ASKING SLIGHTLY INTRUSIVE QUESTIONS IS OKAY IF IT COULD IMPROVE HOW WE WORK
15. OUR VALUES AS AN ENGINEERING TEAM: • COLLECTING DATA IS RISKY SO DO IT CAREFULLY • AVOID DOING MAD SCIENCE
16. THREE CASE STUDIES:
 • ASKING FOR DATA IN THE FIRST PLACE • TRACKING IMMIGRATION STATUS • PREDICTIVE MODELING
17. ETHICAL QUESTIONS TO ASK:
 • IS THIS IN LINE WITH OUR VALUES? • DOES THIS PUT PEOPLE IN HARM’S WAY? • WOULD THIS SURPRISE SOMEONE? • WHAT HAPPENS IF YOU BLOW IT?
18. FIRST: THE POWER DYNAMICS OF ASKING FOR DATA IN THE FIRST PLACE
19. FIRST: ASKING FOR DATA: DOES IT PUT PEOPLE IN HARM’S WAY?
20. FIRST: ASKING FOR DATA: WOULD THIS SURPRISE SOMEONE?
21. FIRST: ASKING FOR DATA: WHAT HAPPENS IF WE BLOW IT?
22. FIRST AND A HALF: THE POWER DYNAMICS OF ASKING FOR DATA IN THE FIRST PLACE (AS A FOR PROFIT BUSINESS)
23. SECOND: TRACKING DANGEROUS DATA LIKE IMMIGRATION STATUS
24. TRACKING IMMIGRATION STATUS BECAUSE IT’S HELPFUL FOR SERVING PATIENTS
25. NOT TRACKING IMMIGRATION STATUS BECAUSE I.C.E. IS A THING NOW
26. SECOND AND A HALF: TRACKING DANGEROUS DATA LIKE IMMIGRATION STATUS (AS A FOR PROFIT BUSINESS)
27. THIRD: THE ALGO AND PREDICTIVE MODELING
28. THIRD AND A HALF: THE ALGO AND PREDICTIVE MODELING (AS A FOR PROFIT BUSINESS)
29. I’M OUT OF STUFF TO TALK ABOUT:
 • PREPARE TO WORK THROUGH ETHICAL QUESTIONS • THINK HARD ABOUT YOUR VALUES AND POWER • CHARGE IN HEADFIRST
30. THANKS •JAMEY, ETHICAL DECISIONS TRACK CURATOR •SOUND PERSON
31. I AM: COLIN FLEMING MY EMAIL: C3FLEMIN@GMAIL.COM MY TWITTER: @RIDETHECATBUS