This book is an intervention -
Chapter 6
Do we need AI or do we need Black feminisms? A poetic guide
By Serena Dokuaa Oduro
Black feminism is rooted in the knowledge that every system and technology has race, gender, and class implications. Black feminism grows from Black women’s experience of living at the intersections of race, gender, class, and many other axes of oppression.
No technology is truly race neutral. Breakthrough research, such as that conducted by Dr Safiya Umoja Noble about the racism and sexism embedded in search engines, has proven it. In fact, the very claim that algorithmic design is race neutral warrants caution in that it is meant to reinforce white, Western, and male norms and power.
New questions must guide the purpose, design, development, and implementation of AI. Not only are new approaches needed to ensure that all groups benefit from AI, they are also needed to curb AI disillusionment. Every report of another instance of racial discrimination in AI further fuels rightful Black scepticism. It is not enough for the goals of AI to be non-discriminatory. AI must seek to benefit and liberate Black women and all historically marginalised groups.
Black feminism provides a roadmap for improving AI, because Black feminist politics “actively [commit] to struggling against racial, sexual, heterosexual and class oppression, and sees as [its] particular task the development of integrated analysis and practice based upon the fact that the major systems of oppression are interlocking.”1 Current research on the impacts of AI ignores those at the intersections of marginalised identities and the differences between historically marginalised groups.2 This poetic guide can be used to measure the usefulness and validity of various AI applications with a Black feminist lens. If an AI system cannot pass the muster of these questions, I recommend seriously considering that it may be AI hype, snake oil, or pseudoscience.
1. Does this algorithm increase individual, interpersonal, and communal well-being for Black women?
An AI system is not useful for me as a Black woman if it reinforces racist and sexist norms. If an AI system perpetuates the hypersexualisaton of Black women, then it does not increase my well-being. If an automated decision system defaults to white content, it does not increase my individual, interpersonal or communal well-being. If the algorithm prioritises content that aligns with colourist, texturist, fatphobic, queerphobic or transphobic norms, then it does not increase individual, interpersonal or communal well-being for Black women.
2. Is this algorithm less biased than a Black feminist?
If the AI system is only better than humans because it is slightly less racist or sexist, then it is not useful for Black women.
3. Should Black feminists be hyped?
If Black feminists are hyped, then I know individual, interpersonal, and communal well-being is going to increase. If Black feminists are hyped, then I know it will be beneficial for historically oppressed people, because Black feminisms are rooted in the knowledge that oppressions are tied together. Since oppressions are tied together, so is liberation.
4. Why is this AI a solution for Black women?
It is typical in the West to look for technical solutions to social problems. But racism, sexism, and class difference cannot be fixed with AI alone. How does this algorithm help Black feminists oppose racial, gender, and class oppression? Who is designing this system? Who is classifying groups?3 Is this AI about bias minimisation or profit maximisation?4 Is this a job for AI or a job that should be given to Black women who most likely already know a solution?
This poetic guide relies on historical evidence that:
Any marginalised group can ask these (and more) questions and it’s important that we do. Current analyses of fairness in AI often flatten the nuance between—and variety in—Black women’s experiences.5 While flattening identities may make it easier for AI, it harms Black women and leads to further AI disillusionment. If the field of AI is to gain public trust, it will have to prove its benefits to all people.
Do we need AI or
Do we need Black feminisms?
Liberation should lead,
Technology should support.
There is no mechanical solution
To sin.
There is only
The purposeful striving towards
Justice.
Serena Dokuaa Oduro is a writer and policy researcher aiming for algorithms to uplift Black communities. She is the Policy Research Analyst at Data & Society.
Notes
1. The Combahee River Collective (1978) A Black Feminist Statement. Capitalist Patriarchy and the Case for Socialist Feminism. Ed. Eisenstein, Z. R. New York: monthly review press.
2. Hanna, A., Denton, E., Smart, A., & Smith-Loud, J. (2020) Towards a critical race methodology in algorithmic fairness. Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency. DOI: 10.1145/3351095.3372826
3. Ibid., at p.10.
4. Benjamin, R. (2019) Race After Technology: abolitionist tools for the new Jim Code. Cambridge, UK: Polity Press. p.30.
5. See note 2, at p.9.
Next: Chapter 7
by James Vincent
Black feminists know
That we
Cannot run from
Oppression.
From work
To the kitchen
Everything
Is political.
No technology is truly race neutral. Breakthrough research, such as that conducted by Dr Safiya Umoja Noble about the racism and sexism embedded in search engines, has proven it. In fact, the very claim that algorithmic design is race neutral warrants caution in that it is meant to reinforce white, Western, and male norms and power.
Black death is
No glitch.
Black erasure is
No glitch.
Black surveillance is
No glitch.
Black false positives is
No glitch.
New questions must guide the purpose, design, development, and implementation of AI. Not only are new approaches needed to ensure that all groups benefit from AI, they are also needed to curb AI disillusionment. Every report of another instance of racial discrimination in AI further fuels rightful Black scepticism. It is not enough for the goals of AI to be non-discriminatory. AI must seek to benefit and liberate Black women and all historically marginalised groups.
I dream of AI being
Crafted by Black hands
And Black dreams.
If I can sit around the table
And gush about AI
With my Mom, Sister
And Aunties, then
I’ll believe in it.
Black feminism provides a roadmap for improving AI, because Black feminist politics “actively [commit] to struggling against racial, sexual, heterosexual and class oppression, and sees as [its] particular task the development of integrated analysis and practice based upon the fact that the major systems of oppression are interlocking.”1 Current research on the impacts of AI ignores those at the intersections of marginalised identities and the differences between historically marginalised groups.2 This poetic guide can be used to measure the usefulness and validity of various AI applications with a Black feminist lens. If an AI system cannot pass the muster of these questions, I recommend seriously considering that it may be AI hype, snake oil, or pseudoscience.
1. Does this algorithm increase individual, interpersonal, and communal well-being for Black women?
An AI system is not useful for me as a Black woman if it reinforces racist and sexist norms. If an AI system perpetuates the hypersexualisaton of Black women, then it does not increase my well-being. If an automated decision system defaults to white content, it does not increase my individual, interpersonal or communal well-being. If the algorithm prioritises content that aligns with colourist, texturist, fatphobic, queerphobic or transphobic norms, then it does not increase individual, interpersonal or communal well-being for Black women.
2. Is this algorithm less biased than a Black feminist?
If the AI system is only better than humans because it is slightly less racist or sexist, then it is not useful for Black women.
3. Should Black feminists be hyped?
If Black feminists are hyped, then I know individual, interpersonal, and communal well-being is going to increase. If Black feminists are hyped, then I know it will be beneficial for historically oppressed people, because Black feminisms are rooted in the knowledge that oppressions are tied together. Since oppressions are tied together, so is liberation.
4. Why is this AI a solution for Black women?
It is typical in the West to look for technical solutions to social problems. But racism, sexism, and class difference cannot be fixed with AI alone. How does this algorithm help Black feminists oppose racial, gender, and class oppression? Who is designing this system? Who is classifying groups?3 Is this AI about bias minimisation or profit maximisation?4 Is this a job for AI or a job that should be given to Black women who most likely already know a solution?
This poetic guide relies on historical evidence that:
If Black women aren’t mentioned, that means Black women will be harmed.
If Black people aren’t mentioned, that means Black people will be harmed.
If gender minorities aren’t mentioned, that means gender minorities will be harmed.
If any historically marginalised group is not mentioned, that group will be harmed.
Any marginalised group can ask these (and more) questions and it’s important that we do. Current analyses of fairness in AI often flatten the nuance between—and variety in—Black women’s experiences.5 While flattening identities may make it easier for AI, it harms Black women and leads to further AI disillusionment. If the field of AI is to gain public trust, it will have to prove its benefits to all people.
Do we need AI or
Do we need Black feminisms?
Liberation should lead,
Technology should support.
There is no mechanical solution
To sin.
There is only
The purposeful striving towards
Justice.
Serena Dokuaa Oduro is a writer and policy researcher aiming for algorithms to uplift Black communities. She is the Policy Research Analyst at Data & Society.
Notes
1. The Combahee River Collective (1978) A Black Feminist Statement. Capitalist Patriarchy and the Case for Socialist Feminism. Ed. Eisenstein, Z. R. New York: monthly review press.
2. Hanna, A., Denton, E., Smart, A., & Smith-Loud, J. (2020) Towards a critical race methodology in algorithmic fairness. Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency. DOI: 10.1145/3351095.3372826
3. Ibid., at p.10.
4. Benjamin, R. (2019) Race After Technology: abolitionist tools for the new Jim Code. Cambridge, UK: Polity Press. p.30.
5. See note 2, at p.9.
Next: Chapter 7
How (not) to blog about an intelligent toothbrush
by James Vincent
