Movatterモバイル変換


[0]ホーム

URL:


There is a fast-growing literature in addressing the fairness ofAI models (fair-AI), with a continuous stream of new concep-tual frameworks, methods, and tools. How much can we trustthem? How much do they actually impact society? We takea critical focus on fair-AI and survey issues, simplifications,and mistakes that researchers and practitioners often under-estimate, which in turn can undermine the trust on fair-AIand limit its contribution to society. In particular, we discussthe hyper-focus on fairness metrics and on optimizing theiraverage performances. We instantiate this observation by dis-cussing the Yule’s effect of fair-AI tools: being fair on averagedoes not imply being fair in contexts that matter. We concludethat the use of fair-AI methods should be complemented withthe design, development, and verification practices that arecommonly summarized under the umbrella of trustworthy AI

Can We Trust Fair-AI?

Ruggieri, Salvatore;Alvarez, Jose M.;Pugnana, Andrea;State, Laura;Turini, Franco
2023

Abstract

There is a fast-growing literature in addressing the fairness ofAI models (fair-AI), with a continuous stream of new concep-tual frameworks, methods, and tools. How much can we trustthem? How much do they actually impact society? We takea critical focus on fair-AI and survey issues, simplifications,and mistakes that researchers and practitioners often under-estimate, which in turn can undermine the trust on fair-AIand limit its contribution to society. In particular, we discussthe hyper-focus on fairness metrics and on optimizing theiraverage performances. We instantiate this observation by dis-cussing the Yule’s effect of fair-AI tools: being fair on averagedoes not imply being fair in contexts that matter. We concludethat the use of fair-AI methods should be complemented withthe design, development, and verification practices that arecommonly summarized under the umbrella of trustworthy AI
2023
Settore INF/01 - Informatica
37th AAAI Conference on Artificial Intelligence
Washington DC, USA
February 7–14, 2023
PROCEEDINGS OF THE 37th AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE
AAAI Press
978-1-57735-880-0
Fair Machine Learning; Fairness Metrics; Yule's Effect
   Artificial Intelligence without Bias
  NoBIAS
  European Commission
  Horizon 2020 Framework Programme
  860630
File in questo prodotto:
FileDimensioneFormato 
CanWeTrustFairAI.pdf

accesso aperto

Tipologia:Published version
Licenza:Non specificata
Dimensione234.41 kB
FormatoAdobe PDF
234.41 kBAdobe PDF

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento:https://hdl.handle.net/11384/136444
Citazioni
  • ???jsp.display-item.citation.pmc???ND
  • Scopus22
  • ???jsp.display-item.citation.isi???ND
  • OpenAlexND
social impact
Conferma cancellazione

Sei sicuro che questo prodotto debba essere cancellato?


[8]ページ先頭

©2009-2025 Movatter.jp