Dec 5, 2023 - Technology

Scoop: Philanthropists offer $30M for more inclusive AI

Illustration of clouds shaped like dollar bill signs

Illustration: Annelise Capossela/Axios

Omidyar Network announced a $30 million investment to prioritize immediate AI risks over existential AI risks and to bridge the gap between AI hype and AI hysteria.

Why it matters: Philanthropists are stepping in to help shift the terms of the AI debate and the Omidyar dollars are intended to broaden who participates in AI discussions.

  • The American public has repeatedly told pollsters they are distrustful of tech giants and concerned about how AI will affect them.
  • But Congress has yet to provide substantial funding, guardrails or a safety net for workers affected by generative AI.
  • An Omidyar Network spokesperson told Axios that its mission is driven by concerns about highly concentrated AI markets and Congress being stuck in learning mode about generative AI.

What's happening: Vice President Kamala Harris last month announced that ten leading foundations would commit more than $200 million towards public interest AI development.

  • The Omidyar Network's cash is the first out the door, and follows Mozilla Foundation publishing its AI funding principles.
  • Mike Kubzansky, CEO of Omidyar Network, told Axios that the funding will foster a "more collective, cross-sector approach to generative AI."

Details: Omidyar Network commissioned focus groups, run by FGS Global in the fall, to help determine what projects it should prioritize.

  • The feedback: immediate harms such as "racial bias, deepfakes and protecting IP" should take priority over existential risk.

Initial partners and projects include AFL-CIO Technology Institute, which will receive funding to bring more stories about worker experiences of AI to Congress and statehouses.

  • Humane Intelligence, led by Rumman Chowdhury and Jutta Williams, will be funded to "scale its AI auditing and evaluation work" which includes red-teaming exercises and global perspectives.
  • The Collective Intelligence Project will be paid to host a series of "AI Alignment Assemblies" — a type of democratic forum designed to surface the views of the whole population on AI.
  • The Economic Security Project and Vanderbilt Policy Accelerator will develop models and proposals for AI governance that include public options for generative AI.

The big picture: AI developers are divided, but seem to be shifting in favor of speeding up rather than slowing down AI development, illustrated by the recent board shake-up at ChatGPT maker OpenAI.

What they're saying: An Omidyar Network representative told Axios that the funding will develop "generative AI infrastructure" which elevates diverse voices in AI development and governance, while promoting innovation and competition.

  • Amanda Ballantyne, director of the AFL-CIO Technology Institute, told Axios that AI can make everyone better off, but only if AI designers work with end users. In AFL-CIO's case that means workers who use enterprise AI.
  • "Workers can tell you pretty quickly what systems will work well, what makes their job safer and easier and more efficient, and what products will require workarounds," Ballantyne said.
Go deeper