Unlocking the Potential: Data as a Medium for Design & Justice
0 likes52 views
As a designer and data scientist, I work with data in service of meeting peoples needs. This data is inherently subjective, which is what makes it both an excellent medium for design and vulnerable to misuse. Ill share what it means to design with data and how data science can contribute to and augment the design process.Once we see whats exciting and newly possible, well delve into why now, more than ever, human-centered design matters as we discuss ethics and the impact of AI designs on individuals, communities, and societies. Talk given by Jess Freaner (IDEO) at UX Strategy Meetup in Chicago - November 2019
![Gender bias in
seatbelts
For millennia, [design]
has functioned on the
assumption that male
bodies can represent
humanity as a whole.
- Caroline Criado-P辿rez
Invisible Women: Data bias in a world
designed for men.](https://image.slidesharecdn.com/ideouxstrategymeetup2019-191126220620/85/Unlocking-the-Potential-Data-as-a-Medium-for-Design-Justice-36-320.jpg)
Unlocking the Potential: Data as a Medium for Design & Justice
- 1. Unlocking the Potential DATA AS A MEDIUM FOR DESIGN & JUSTICE 11.19.2019 UNLOCKINGTHEPOTENTIAL UX Meetup
- 2. Jess Freaner Senior Design Lead & Data Scientist, IDEO MY POV: Data ethics, human-centered design of intelligent products, services & systems BACKGROUND: BA in Psychology, co-organizer of Chicago Data Science + Chicagos Women in Machine Learning & Data Science (WiMLDS)
- 5. UNLOCKING THE POTENTIAL SAN FRANCISCO PALO ALTO CHICAGO CAMBRIDGE LONDON MUNICH SHANGHAI TOKYO NEW YORK NATIONALDESIGNCONFERENCE2019
- 6. Founded in 1991, Chicago studio has 90+ designers of di鍖erent backgrounds including researchers, artists, and data scientists. UNLOCKING THE POTENTIAL IDEO Chicago
- 7. ORGANIZATION DESIGNER INDUSTRIAL DESIGNER DESIGN RESEARCHER INTERACTION DESIGNER SYSTEMS DESIGNER CODER PHYSICIAN PROTOTYPER MECHANICAL ENGINEER COMMS DESIGNER ARCHITECT WRITER FOOD SCIENTIST MANUFACTURING SPECIALIST BUSINESS DESIGNER INFORMATION TECHNOLOGY HUMAN FACTORS DATA SCIENTIST
- 8. Data Science is my design discipline. UNLOCKINGTHEPOTENTIAL
- 10. Signals in the world that people decided to capture & quantify. UNLOCKINGTHEPOTENTIAL
- 19. Data is a human construct. UNLOCKINGTHEPOTENTIAL
- 20. Data is not truth. UNLOCKINGTHEPOTENTIAL
- 21. DESIGNFORAUGMENTEDINTELLIGENCE 21 Data is a medium for design.
- 23. As new capabilities emerge, new potential for how we can meet peoples needs emerge. UNLOCKINGTHEPOTENTIAL
- 24. Augmented Intelligence Machine + Human UNLOCKINGTHEPOTENTIAL
- 25. UNLOCKINGTHEPOTENTIAL The design process gets us there.
- 27. The design process UNLOCKINGTHEPOTENTIAL ASKING QUESTIONS BUILDING CONFIDENCE CONTINUOUS STORYTELLING
- 30. Prototyping: building to learn & think UNLOCKINGTHEPOTENTIAL
- 33. There is subjectivity, intuition, choice & creativity in all three. UNLOCKINGTHEPOTENTIAL ASKING QUESTIONS BUILDING CONFIDENCE CONTINUOUS STORYTELLING
- 34. With all the subjectivity, biases seep in. UNLOCKINGTHEPOTENTIAL
- 35. This is true of all design. UNLOCKINGTHEPOTENTIAL
- 36. Gender bias in seatbelts For millennia, [design] has functioned on the assumption that male bodies can represent humanity as a whole. - Caroline Criado-P辿rez Invisible Women: Data bias in a world designed for men.
- 37. Gender bias in seatbelts The seatbelt is designed in such a way that a woman in a car crash is 17 percent more likely to die than a man.
- 38. Racial bias in photography: Shirley Cards Photography is not just a system of calibrating light, but a technology of subjective decisions. -Sarah Lewis Harvard University
- 39. Our data re鍖ects priorities of the powerful. NATIONALDESIGNCONFERENCE2019
- 40. Sometimes data cannot and should not speak on its own. - Catherine DIgnazio & Lauren Klein co-authors of Data Feminism Consider the context* *One of the 7 central tenets of Data Feminism by Catherine DIgnazio & Lauren Klein
- 41. 99 Men + 1 Woman Sometimes power makes you more visible
- 42. Predictive policing and sentencing algorithms used in many US cities rely on data gathered disproportionally about some populations. Sometimes power means invisibility Machine Bias by ProPublica 2016
- 43. Colleges with higher sexual assault numbers can re鍖ect the shifting incentives and increasing support structures for survivors. As visibility shifts so does our data
- 44. When working with data, being aware of this vulnerability is especially important. UNLOCKINGTHEPOTENTIAL
- 45. When working with data, designing for this vulnerability is especially important. UNLOCKINGTHEPOTENTIAL
- 46. UNLOCKINGTHEPOTENTIAL Why design matters with data Veneer of objectivity. Perpetuation of prior biases. Scale of unintended consequences.
- 47. When it has gone wrong SELF-PERPETUATION OF MALE INFLUENCE MISINTERPRETATION OF SEXUAL ASSAULT STATS NATIONALDESIGNCONFERENCE2019 99 Men + 1 Woman RACIAL DISPARITY IN PREDICTIVE POLICING
- 48. These examples are not the failures of algorithms UNLOCKINGTHEPOTENTIAL
- 49. They are failures of design. UNLOCKINGTHEPOTENTIAL
- 50. We can do better. UNLOCKINGTHEPOTENTIAL
- 52. Data as a stakeholder HOW IS IT CREATED? WHOS & WHATS MISSING?HOW IS IT BEING USED? UNLOCKINGTHEPOTENTIAL
- 54. UNLOCKINGTHEPOTENTIAL Capturing minimum viable data Limit the amount of data to gather and retain. Do you need to sense who is in a room, or simply that someone is there? Do you need to know what someone said, or just the tone and volume of their voice?
- 55. UNLOCKINGTHEPOTENTIAL But you cant anticipate everything, so how do you respond?
- 58. UNLOCKINGTHEPOTENTIAL City of Chicago (in some ways) Chicagos Array of Things (AoT) https://arrayofthings.github.io/
- 59. https://www.ajlunited.org/ UNLOCKINGTHEPOTENTIAL Algorithmic Justice League Research by Joy Buolamwini, MIT founder of AJL
- 60. UNLOCKINGTHEPOTENTIAL Data Feminism Data is power, and that power is wielded unequally - Catherine DIgnazio & Lauren Klein https://is.gd/datafem
- 61. UNLOCKINGTHEPOTENTIAL IDEOs data principles Data is Not Truth Dont Presume the Desirability of Intelligent Systems Respect Privacy and the Collective Good Unintended Consequences of AI are Opportunities for Design
- 65. Whats a better font or #creativetensions Papyrus? Comic sans?
- 67. Does data lead to or #creativetensions Questions? Answers?
- 68. Does data help or #creativetensions Personalize? Homogenize?
- 69. To reduce bias & harm caused by AI, we need or #creativetensions More voices at the table More checks & balances
- 70. Methods used to measure algorithm goodness should be or #creativetensions Equitable Objective
- 71. The ethical concerns that come along with AI are or #creativetensions A whole new world Same s***, different day