Discussion: 300 words

Ethics and Professionalism in CS I N F O R M E D C O N S E N T

1

Human Subjects Research A scientific investigation that involves humans as the test subjects.

Benefits and harm balance.

2

Tuskegee Case (1) Done in the period between 1932-1972 in Macon Alabama by the US public health service.

An observational study to monitor the development of untreated Syphilis.

600 (399 with, 201 without) African-American sharecroppers were involved in the study.

They were only told that they are receiving free health-care and hot meals.

3

Tuskegee Case (2) In 1947, Penicillin became a standard treatment for Syphilis. ◦ Poor subjects were never informed!!!

A public outcry halted this study in 1972. ◦ The harm caused on subjects not being treated outweighs the medical value out of

this study. ◦ More harm than benefit

4

The Tuskegee case?

5

Informed Consent Human subjects MUST ◦ Be informed about the experiment ◦ Consent to the experiment without any

type of coercion. ◦ Have the right to withdraw at any time.

Researchers shouldn’t (INTENTIONALLY) cause any harm to subjects.

Required by federal law. ◦ ONLY for government funded studies!!

6

Institutional Review Board (IRB) It is sometimes hard for subjects to assess the benefits vs harm in an experiment they are involved in. ◦ Experimenters cannot do that alone!

Benefits vs harm should be assessed by whom?

IRBs comprises diverse members ◦ Including non-scientists.

It balances the potential harm on human subjects against benefits to the society.

Manages any violations of informed consent

7

Informed Consent (Always possible??)

8

Business studies Informed consent and IRB review are NOT required if the study is done by businesses.

A/B testing for online companies ◦ Group of customers shown version A ◦ Other group shown version B ◦ Compare

NOT considered human subjects research!!!! ◦ Considered a good business practice.

9

Facebook/Cornell experiment In 2012, researchers at Facebook and Cornell university conducted an experiment to manipulate user mood (Emotional Contagion). ◦ Published in 2014 ◦ News feed of some Facebook users (700, 000)

were manipulated ◦ Some were shown positive articles ◦ Others were shown negative articles

◦ Results were really interesting!! ◦ Mood is contagious!

Was it legal? Was it ethical?

10

Facebook / Cornell Experiment

11

OKCupid case Online matching company. ◦ Match couple profiles ◦ A matching algorithm ◦ Compatibility score

The love is blind day (Suppressed photos) ◦ This results in more conversations

They tinkered with the algorithm ◦ They told couples they were compatible

even if they are not!!!! ◦ And vice versa ◦ Results were interesting!!!

12

OKCupid case (2)

13

Experiment

Compatible Not Compatible

Truth Compatible Good Good

Not Compatible Good Bad

OKCupid

14