We here So excited for this https://twitter.com/_w0bb1t_/status/1335666598200434689

@ResistanceAI #NeurIPS2020

"AI-adjacent privilege" #ResistanceAI
This workshop is EVERYTHING

Thank you for this @ResistanceAI

And @ResistanceAI now has a panel with literally so many of my favourite people in the ML fields @timnitGebru @Abebab @vilarinoramon @red_abebe & @tetisheri

"Is AI trapped in a paradigm that pursues efficiency?" ~ @vilarinoramon

#NeurIPS2020

@Abebab talks about her recent work on what ML values are most prominent in research?

Accuracy
Performance
Efficiency

These values are driving Dev and deployment of ML systems.

And @red_abebe brings up "Efficiency for whom?"

She highlights how we can redefine the definition of efficiency. They're newly defined and we can change them

Next @timnitGebru refers to @Trevornoah 's Born a Crime, where the Apartheid government tracked data in depth marginalized people. She highlights how the machine is well oiled

@tetisheri highlighting the "strange, interesting and perverse clash of the efficiency paradigm and the social good intent" in the global south

"As if technology will leapfrog the continent out of poverty.. as if by magic. With that narrative comes this damaging view of Africa" ~ @Abebab

And @vilarinoramon talking about the right to deny this reality. Pointing to "Ideas to postpone the end of the world" by Ailton Krenak coming out in English soon.

First @timnitGebru says she used to want everyone to have access to AI around the world. However the paradigm is already Western and they're trying to push this way of thinking

Shout-out to @MasakhaneNLP and @DeepIndaba from @timnitGebru

. @red_abebe highlights how it is easier for someone from outside the continent to access their data than for Africans to access their own data

And @Abebab is highlighting that despite being in western institute @MD4SG are try to get at the nuances of exposing power inequalities

And @tetisheri talking about the political situation in India, acknowledging that the definition of Decolonisation differs from place to place.

More general desire: If AI has so much public impact,..., it should be recognised as a public science.

. @tetisheri leaves us with the truth that when you approach interdisciplinary groups to involve them in the process, it's very intimidating - the field needs to be demystified

@timnitGebru : "Even after my treatment, I still believe there is some space for people to do this work in these companies. But we need to make sure they can do it"

And @red_abebe talks of resistance as an ecosystem. "we can afford to lose more people in an attempt to resist".

Resistance can come at varying levels of cost. Expressing gratitude to @timnitGebru and the cost she felt by her resistance

According to @tetisheri, it's time to start bring knowledge sharers who have experience organising against authoritarian governments into the ML community