data-hazards
data-hazards copied to clipboard
very-good-science
Things to Read
This is a list of reading that we think may be relevant to read/cite when we write up the Data Hazards project into a paper. If you think there's anything else we should be reading, please let us know in a comment below:
- [ ] Innovating Like an Optimist, Preparing Like a Pessimist: Ethical Speculation and the Legal Imagination by Casey Fiesler (2021)
Other ethics toolkits, checklists, question lists, etc:
- Association For Computational Linguistics conference ethics review questions
- UK Statistics Authority Public Good Checklist
- UK Statistics Authority Ethics Self-Assessment tool
Other people-centric ways of identifying ethics issues
"Research Ethics By Design" (2017) https://journals.sagepub.com/doi/pdf/10.1177/1747016116673135 Really helpful framing of the meaning of ethics by design. Our work builds this as a framework for applying it in real life.
Mitigating dataset harms requires stewardship: Lessons from 1000 papers https://arxiv.org/pdf/2108.02922.pdf
Makes recommendations about how to make dataset use more ethical -->
(8.1 Dataset creators): Make ethically-salient information clear and accessible
(8.4 Dataset users and other researchers): Our findings show how a dataset’s impact is not fully understood at the time of its creation. We recommend that the community systematize the retrospective study of datasets to understand their ethical implications, potential shortcomings, and misuse. Researchers should not wait until the problems become serious and there is an outcry.
Reviews/overviews of ethics AI frameworks:
- The ethics of algorithms: Mapping the debate
- From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices
https://dl.acm.org/doi/pdf/10.1145/3479582 Boyd KL. Datasheets for Datasets help ML Engineers Notice and Understand Ethical Issues in Training Data. Proceedings of the ACM on Human-Computer Interaction. 2021 Oct 18;5(CSCW2):1-27.
And of course the original Datasheets for Datasets: https://www.microsoft.com/en-us/research/uploads/prod/2019/01/1803.09010.pdf
https://www.uu.nl/en/news/utrecht-data-school-develops-data-ethics-decision-aid Utrecht Data School's Data Ethics Decision Aid
https://www.adalovelaceinstitute.org/feature/prototyping-ai-ethics-futures/ May 2021 "A week-long series of dialogues and workshops highlighting the new possibilities of a humanities-led, broadly engaging approach to data and AI ethics"
https://weallcount.com/tools/ some very relevant tools and the blogs are brilliant
Sorry - not sure how to add links but here are the urls. https://www.canada.ca/en/government/system/digital-government/digital-government-innovations/responsible-use-ai/algorithmic-impact-assessment.html https://www.adalovelaceinstitute.org/report/algorithmic-impact-assessment-case-study-healthcare/ https://datanutrition.org/
Sorry again - should have introduced myself. I was attending your workshop this morning and saw this issue and wanted to contribute. Hope that is OK.
Z-inspection: http://z-inspection.org/wp-content/uploads/2021/12/HowtoAssessTrustworthyAI.pdf