Radical AI Network

What is radical AI       Radical AI principles       Radical AI work       The Radical AI Network

Radical AI Principles

  1. We recognize that power is distributed unevenly, and People of Color, Black, Indigenous, Women, Queer, Trans, Gender Non-Conforming, Poor, Dis-abled, and other communities are pushed to the margins. We commit to resisting these interlocking systems of oppression, including Racism, Coloniality, Casteism, Heteropatriarchy, Capitalism, and Ableism. We commit to striving toward collective liberation.1
  2. We center radical work by communities at the margins.
  3. We recognize that all technologies rearrange power.
  4. We are critical of how AI shifts power. In particular, we recognize AI is frequently extractive, exploitative, surveilling, controlling, prescriptive, and reductionist. We recognize AI frequently prevents consent, deliberation, investigation, intervention, resistance, and agency.
  5. We recognize the privilege of being adjacent to institutions creating technologies. We use this privilege to step up to expose oppressive technologies and to center those harmed, and to step back, ceding power, listening, and co-conspiring when those voices speak.
  6. We value dreaming up and building human/AI systems that put power in the hands of the people. We recognize this is vulnerable and important work.
  7. We recognize that reforming harmful AI can reduce harm to some while legitimizing (“ethics-washing”) systems still harming others2. Only by understanding the root of the harm and centering those most impacted can we make these complex tactical decisions, deciding when to minimize harm done by AI and when to call for its abolition3. We recognize that no one has a ‘view from nowhere’4. We practice reflexivity, naming who we are, how who we are influences what we do, and whose perspectives are missing.

  1. This principle is informed by The Combahee River Collective Statement ↩

  2. This tension is prominent in efforts such as ‘AI for Social Good,’ ‘AI Ethics,’ and ‘Algorithmic Fairness, Accountability, and Transparency’. These are not inherently Radical AI objectives: Creating an AI, for example, to decide whether to jail someone in a way that is “Fair,” “Accountable,” and “Transparent” (a common motivation for work in AI for Social Justice areas) doesn’t inherently question the Prison Industrial Complex it is created for and the shifts power to the carceral system. Radical AI is a response to these and similar AI for Social Justice efforts: Radical AI moves to recenter the affected marginalized communities and to ask how these technologies redistribute power in the world. ↩

  3. This principle is informed by Audre Lorde’s The Master’s Tools Will Never Dismantle the Master’s House ↩

  4. The “view from nowhere” is developed in Donna Haraway’s Situated Knowledges ↩

Get Involved

If you are doing Radical AI work, you are part of the Radical AI project! If you would like to join the Radical AI Network — a community of self-identified Radical AI folks connecting, learning, and deepening our radical AI work — we organize ourselves in a Slack workspace. To join, email a little about yourself to the.radical.ai.network@gmail.com.