Jump to content

Draft:Public AI

From Wikipedia, the free encyclopedia

Public AI refers to artificial intelligence (AI) systems designed and governed as public infrastructure, especially those that prioritize universal access, public accountability, and long-term sustainability.[1][2] In practice, public AI can refer to layers of the AI stack that are owned and operated by public sector entities, to private AI companies regulated like public utilities, or to certain forms of public-private partnership.[3] Examples include public AI models such as AI Sweden's GPT-SW3[4] and AI Singapore's SEA-LION[5], AI-focused public compute projects such as New York State's Empire AI[6] and the National AI Research Resource (NAIRR)[7], as well as proposed public data banks for AI training[8].

References

[edit]
  1. ^ "Public AI Network". Public AI Network. Retrieved 2024-10-18.
  2. ^ "Creating a public counterpoint for AI | The Mozilla Blog". blog.mozilla.org. Retrieved 2024-10-18.
  3. ^ Sitaraman, Ganesh; Pascal, Alex (2024-09-29). "The National Security Case for Public AI" (PDF). Vanderbilt Policy Accelerator.
  4. ^ "GPT-SW3 | AI Sweden". www.ai.se. Retrieved 2024-10-18.
  5. ^ "SEA-LION - AI Singapore". 2023-11-20. Retrieved 2024-10-18.
  6. ^ "Empire AI | Responsible AI Innovation". Empire AI. Retrieved 2024-10-18.
  7. ^ "National Artificial Intelligence Research Resource Pilot | NSF - National Science Foundation". new.nsf.gov. Retrieved 2024-10-18.
  8. ^ "Unlocking AI for All: The Case for Public Data Banks". Default. Retrieved 2024-10-18.