Successfully reported this slideshow.

AI and Social Justice: From Avoiding Harms to Positive Action

0

Share

Loading in …3
×
1 of 18
1 of 18

AI and Social Justice: From Avoiding Harms to Positive Action

0

Share

Download to read offline

Talk at The AI Summit New York, 8th Dec. 2021.

https://www.alandix.com/academic/talks/AI-Summit-NY-2021-AISJ/

AI and in particular large data machine learning are transforming many areas of society including healthcare, education, and finance. At their best these offer the potential to improve society, for example, finding new pharmaceuticals. However, they may also reproduce or reinforce existing divisions and inequalities as well as creating new problems. This has been evident in high-profile news items such as the case of racial discrimination in facial recognition systems used in policing. However, some of the deepest problems in unequal access to technology are still to be fully felt.

Fortunately, AI can also be used positively to address issues of social injustice, for example human-rights organisations scanning public domain images for evidence of abuse, or software using adversarial techniques to reduce bias in training data. At best some companies and institutions are addressing these issues proactively, seeking ways to ensure they prevent or detect problems before they happen, for others this may be a rear-guard action to fix problems that have already emerged.

In this talk we will present a high-level landscape of the ways on which AI interacts with social justice and illustrate this through examples so that we can take positive action for a fairer world.

Talk at The AI Summit New York, 8th Dec. 2021.

https://www.alandix.com/academic/talks/AI-Summit-NY-2021-AISJ/

AI and in particular large data machine learning are transforming many areas of society including healthcare, education, and finance. At their best these offer the potential to improve society, for example, finding new pharmaceuticals. However, they may also reproduce or reinforce existing divisions and inequalities as well as creating new problems. This has been evident in high-profile news items such as the case of racial discrimination in facial recognition systems used in policing. However, some of the deepest problems in unequal access to technology are still to be fully felt.

Fortunately, AI can also be used positively to address issues of social injustice, for example human-rights organisations scanning public domain images for evidence of abuse, or software using adversarial techniques to reduce bias in training data. At best some companies and institutions are addressing these issues proactively, seeking ways to ensure they prevent or detect problems before they happen, for others this may be a rear-guard action to fix problems that have already emerged.

In this talk we will present a high-level landscape of the ways on which AI interacts with social justice and illustrate this through examples so that we can take positive action for a fairer world.

More Related Content

Related Books

Free with a 14 day trial from Scribd

See all

AI and Social Justice: From Avoiding Harms to Positive Action

  1. 1. AI and Social Justice Clara Crivellaro Open Lab, Newcastle University, Newcastle, UK Alan Dix Computational Foundry, Swansea University, Wales, UK https://alandix.com/ai4sj/ @alanjohndix @unsolopunto
  2. 2. AI … broad (e.g. EU defn.) big data statistics ML and NN classic AI
  3. 3. Social Justice equity fairness empowerment
  4. 4. AI is embedded in society US judges’ parole decisions China mass face recognition Autonomous weapons Loan and insurance decisions France was ruled by AI in 2019!
  5. 5. ACCESS DOING IT RIGHT POLICY
  6. 6. ACCESS DOING IT RIGHT POLICY
  7. 7. Avoid harm • Bans • Inclusive data sets • Auditing • Base rates and correlates • Neutral is not unbiased
  8. 8. ACCESS DOING IT RIGHT POLICY DOING IT RIGHT
  9. 9. UK A’level ‘mutant algorithm’ • Entrenching inequality • Tried to be fair • Lack of data • Lack of trust • Unreasonable expectations of algorithms • Mirror on existing system?
  10. 10. context and system AI Richard Foster CC BY-SA 2.0 https://en.wikiversity.org/wiki/File:Different_people.jpg David Iliff CC BY-SA 3.0 https://en.wikipedia.org/wiki/File:Hopetoun_falls.jpg Michael Janich CC BY-SA 3.0 https://commons.wikimedia.org/wiki/File:Hongkong_central_kowloon-full.jpg
  11. 11. ACCESS DOING IT RIGHT POLICY
  12. 12. Human Rights Information and Documentation Systems https://huridocs.org/ https://www.uwazi.io/
  13. 13. AI and Social Justice Clara Crivellaro Open Lab, Newcastle University, Newcastle, UK Alan Dix Computational Foundry, Swansea University, Wales, UK https://alandix.com/ai4sj/ @alanjohndix @unsolopunto

Editor's Notes

  • National Archives Identifier (NAID) 542069. https://commons.wikimedia.org/wiki/File:Martin_Luther_King_-_March_on_Washington.jpg
    https://en.wikipedia.org/wiki/File:Official_program_-_Woman_suffrage_procession_March_3,_1913_-_crop.jpg
  • Richard Foster CC BY-SA 2.0 https://en.wikiversity.org/wiki/File:Different_people.jpg
    David Iliff CC BY-SA 3.0 https://en.wikipedia.org/wiki/File:Hopetoun_falls.jpg
    Michael Janich CC BY-SA 3.0 https://commons.wikimedia.org/wiki/File:Hongkong_central_kowloon-full.jpg

  • https://www.forbes.com/sites/carltonreid/2021/11/06/bidens-12-trillion-infrastructure-bill-hastens-beacon-wearing-for-bicyclists-and-pedestrians-to-enable-detection-by-connected-cars/
    https://upload.wikimedia.org/wikipedia/commons/thumb/e/ee/Cyclist_recovers_after_traffic_accident.jpg/1024px-Cyclist_recovers_after_traffic_accident.jpg
  • https://commons.wikimedia.org/wiki/File:Tinkoff_Bank_App.jpg
  • https://huridocs.org/2020/06/huridocs-team-members-reflect-on-google-dot-org-fellows-collaboration/
  • ×