There has been some great work on framing AI ethics issues as ultimately about power.
I want to elaborate on *why* power imbalances are a problem. 1/
I want to elaborate on *why* power imbalances are a problem. 1/
*Why* power imbalances are a problem:
- those most impacted often have least power, yet are the ones to identify risks earliest
- those most impacted best understand what interventions are needed
- often no motivation for the powerful to change
- power tends to be insulating 2/
- those most impacted often have least power, yet are the ones to identify risks earliest
- those most impacted best understand what interventions are needed
- often no motivation for the powerful to change
- power tends to be insulating 2/
The Participatory Approaches to ML workshop at #ICML2020 was fantastic. The organizers highlighted how even many efforts for fairness or ethics further *centralize power*
https://twitter.com/math_rachel/status/1284168523841847297">https://twitter.com/math_rach... 3/
https://twitter.com/math_rachel/status/1284168523841847297">https://twitter.com/math_rach... 3/
I think leaders at major tech companies often see themselves as benevolent & believe that critics do not understand the issues as well as they do. E.g. Facebook criticizes Kevin Roose& #39;s data as inaccurate, yet won’t share other data:
https://twitter.com/EthanZ/status/1328800197699375104">https://twitter.com/EthanZ/st... 4/
https://twitter.com/EthanZ/status/1328800197699375104">https://twitter.com/EthanZ/st... 4/
Only the issues *most visible* to system designers get addressed. With this model, many issues are not prioritized until they are incredibly widespread & causing serious harm (as opposed to addressing earlier). 5/ https://twitter.com/math_rachel/status/1283249233613619200?s=20">https://twitter.com/math_rach...
A classic example of tech platforms being insulated from the harms caused is the Black women who raised an alarm on deceptive sock-puppets and coordinated harassment in 2014, yet Twitter failed to respond 6/ https://slate.com/technology/2019/04/black-feminists-alt-right-twitter-gamergate.html">https://slate.com/technolog...
What motivates the powerful to act: I still think about Facebook hiring 1,200 content moderators in Germany in < 1 year to avoid a hefty fine, vs after 5 years of warnings about genocide in Myanmar, just hiring “dozens” of Burmese content moderators 7/ https://twitter.com/math_rachel/status/1021906783944597505">https://twitter.com/math_rach...
As more background, here is a thread of work on framing AI ethics issues as being about power 8/ https://twitter.com/math_rachel/status/1284976543769309184">https://twitter.com/math_rach...
Machine learning often has the effect of centralizing power 9/ https://twitter.com/math_rachel/status/1293692877298532352">https://twitter.com/math_rach...
Also, related: Tech is not neutral 10/ https://twitter.com/math_rachel/status/1274908170901569537">https://twitter.com/math_rach...
Question: What reasons or resources have you found helpful in convincing those that don’t see power imbalances within tech & AI as a problem, believing the solution is primarily for those with power to wield it more benevolently? 11/