Elamur Research⌕
☰
home / papers / 2022 / Ethics Sheet for Automatic Emotion Recognition and Sentiment Analysis

Ethics Sheet for Automatic Emotion Recognition and Sentiment Analysis

Paper · 2022 · Saif M. Mohammad

Summary

Emotion recognition and sentiment analysis tools can deliver real benefits, for example in health care or commerce. The same technology can also be used for surveillance, political manipulation, or exploiting vulnerable people. Mohammad addresses the lack of a single organized reference on the ethics of this field by writing an "ethics sheet" for automatic emotion recognition (AER). This is a semi-standardized format he proposed earlier, similar to a survey but centered on ethical issues.

The sheet draws on the AI ethics and emotion-recognition literatures, discussions with scholars, and public feedback on a blog draft. It first describes input modalities, the different ways the task is framed, and typical applications. It then sets out fifty ethical considerations grouped by theme. These expose hidden assumptions in task framing, data, methods, and evaluation, with special focus on privacy and on effects for social groups. An appendix condenses them into practical recommendations.

The goal is to prompt careful thought before systems are built. The document also serves as an introduction to AER and a template for other NLP tasks. The author presents it as one perspective, not a definitive standard.

Key points

  • The main sheet (version 1.0) is divided into four sections: one on input modalities and scope, one defining the task, one on applications, and one on ethical issues. It focuses on text-based AER in NLP, but it also lists other modalities: facial expressions, gait, physiological signals, touch, and speech.
  • The author argues machines can never determine a person's true emotional state, only limited aspects of it, in a way he compares to machine translation or web search. He separates three framings: emotions felt, emotions perceived by others, and emotions the speaker intends to convey. The "felt" framing is the one most often misinterpreted.
  • For privacy, the sheet recommends Ann Cavoukian's seven principles of privacy by design, consent that scales with data sensitivity, anonymization or differential privacy, and permission for secondary use. It also warns about mass-surveillance threats to rights such as protest and free expression.
  • For social groups, it recommends diverse annotators, reporting annotator demographics, and possibly separate label sets per demographic group instead of a single majority vote. Results should be evaluated disaggregated by group, and intersectionality should be considered.
  • Further recommendations cover explainability planned from the start, contestability mechanisms, and visualizations that let users drill down to the source data. It also asks authors to report carbon footprint, cost per example, training-set size, hyperparameter count, and budget–accuracy curves.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🏛️National Research Council CanadaOttawa · Canada40 works · 20,943 citations

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksSocial Sciences · 744 worksPsychology · 798 worksExperimental and Cognitive Psychology · 463 worksSafety Research · 221 works

🧭 Topics 2Research subjects and languages the work is about.Sentiment & emotion · 608 worksBias, fairness & safety · 209 works

Cite

Cite as Saif M. Mohammad. 2022. Ethics Sheet for Automatic Emotion Recognition and Sentiment Analysis. Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.1162/coli_a_00433

@article{mohammad2022ethics,
  title = {{Ethics Sheet for Automatic Emotion Recognition and Sentiment Analysis}},
  author = {Saif M. Mohammad},
  journal = {Computational Linguistics},
  year = {2022},
  doi = {10.1162/coli\_a\_00433},
  url = {https://doi.org/10.1162/coli_a_00433}
}

References

44 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Agrawal, Aishwarya, Dhruv Batra, and Devi Parikh. 2016. Analyzing the behavior of visual question answering models. arXiv preprint arXiv:1606.07356. https://doi.org/10.18653/v1/D16-1203 AI2. 2019. Crowdsourcing: Pricing ethics and best practices. Medium. https://medium .com/ai2-blog/crowdsourcing-p… ↗
  2. Pew Research Center. Bagby, R. Michael, James D. A. Parker, and Graeme J. Taylor. 1994. The twenty-item Toronto Alexithymia scale: I. Item selection and cross-validation of the factor structure. Journal of Psychosomatic Research, 38(1):23–32. https://doi.org/10.1016 /0022-3999(94)90005-1 Bamberg, Mi… — [doi:10.1016/s0388-0001(97)00004-1](https://doi.org/10.1016/s0388-0001%2897%2900004-1) find
  3. Houghton Mifflin Harcourt. Barrett, Lisa Feldman. 2017b. The theory of constructed emotion: An active inference account of interoception and categorization. Social Cognitive and Affective Neuroscience, 12(1):1–23. https://doi.org/10.1093/scan/nsx060 Barrett, Lisa Feldman, Ralph Adolphs, Stacy Marse… — [doi:10.1093/scan/nsx060](https://doi.org/10.1093/scan/nsx060) ↗
  4. Basile, Valerio. 2020. It’s the end of the gold standard as we know it. On the impact of pre-aggregation on the evaluation of highly subjective tasks. 2020 AIxIA Discussion Papers Workshop, AIxIA 2020 DP, volume 2776, pages 31–40. find
  5. Bauer, Greta R., Jessica Braimoh, Ayden I. Scheim, and Christoffer Dharma. 2017. Transgender-inclusive measures of sex/gender for population surveys: Mixed-methods evaluation and recommendations. PLOS ONE, 12(5):e0178043. https://doi.org/10 .1371/journal.pone.0178043 Bender, Emily M., Timnit Gebru,…
  6. Artificial Intelligence, music recommendation, and the curation of culture. Technical report. Schwartz Reisman Institute. Boyle, Gregory J. 1995. Myers–Briggs type indicator (MBTI): Some psychometric limitations. Humanities & Social Sciences Papers, 30(1):71–74. https://doi.org/10 .1111/j.1742-9544.… — [doi:10.1108/jet-02-2017](https://doi.org/10.1108/jet-02-2017) find
  7. Chomsky, Noam. 1975. Reflections on Language. find
  8. Pantheon. Chomsky, Noam. 2014. Aspects of the Theory of Syntax, volume 11. find
  9. MIT Press. Cobb-Clark, Deborah A. and Stefanie Schurer. 2012. The stability of big-five personality traits. Economics Letters, 115(1):11–15. https://doi.org/10.1016/j.econlet.2011.11.015 De Choudhury, Munmun, Michael Gamon, Scott Counts, and Eric Horvitz. 2013. Predicting depression via social medi… ↗
  10. De Cicco, Roberta, Riccardo Palumbo, et al. 2020. Should a chatbot disclose itself? Implications for an online conversational retailer. In International Workshop on Chatbot Research and Design, pages 3–15. https://doi.org/10.1007/978-3-030 -68288-0 1 Denton, Emily, Alex Hanna, Razvan Amironesei, And… — [doi:10.1007/978-3-030](https://doi.org/10.1007/978-3-030) ↗
  11. Dickson, Ben. 2018. Why AI must disclose that it’s AI. PC Magazine. https://www.pcmag.com/opinions/why-ai-must -disclose-that-its-ai Dolmaya, Julie McDonough. 2011. The ethics of crowdsourcing. Linguistica Antverpiensia, New Series–Themes in Translation Studies, 10:97–110. ↗
  12. Dwibhasi, Sharat, Dheeraj Jami, Shivkanth Lanka, and Goutam Chakraborty. 2015. Analyzing and visualizing the sentiments of Ebola outbreak via tweets. In Proceedings of the SAS Global Forum, pages 26–29. find
  13. Eichstaedt, Johannes C., Hansen Andrew Schwartz, Margaret L. Kern, Gregory Park, Darwin R. Labarthe, Raina M. Merchant, Sneha Jha, Megha Agrawal, Lukasz A. Dziurzynski, Maarten Sap, Christopher Weeg, Emily E. Larson, Lyle H. Ungar, and Martin E. P. Seligman. 2015. Psychological language on Twitter p… — [doi:10.1177/0956797614557867](https://doi.org/10.1177/0956797614557867) find
  14. Oxford University Press. Ervin-Tripp, Susan. 1973. Some strategies for the first two years. In Cognitive Development and Acquisition of Language, pages 261–286. https://doi.org/10.1016/B978-0-12-505850-6.50018-9 Fletcher-Watson, Sue, Jon Adams, Kabie Brook, Tony Charman, Laura Crane, James Cusack,… ↗
  15. Floridi, Luciano. 2014. Open data, data protection, and group privacy. Philosophy & Technology, 27(1):1–3. https://doi.org/10.1007/s13347-014-0157-8 Fort, Karën, Gilles Adda, and K. Bretonnel Cohen. 2011. Last words: Amazon Mechanical Turk: Gold mine or coal mine? Computational Linguistics, 37(2):… — [doi:10.1007/s13347-014-0157-8](https://doi.org/10.1007/s13347-014-0157-8) ↗
  16. Fraser, Kathleen C., Frauke Zeller, David Harris Smith, Saif M. Mohammad, and Frank Rudzicz. 2019. How do we feel when a robot dies? Emotions expressed on Twitter before and after hitchBOT’s destruction. In Proceedings of the Tenth Workshop on Computational Approaches to Subjectivity, Sentiment and… find
  17. Google. 2018. Google AI Blog. https://ai.googleblog.com/2018/09/the-what-if -tool-code-free-probing-of.html Grant, Adam. 2013. Say goodbye to MBTI, the fad that won’t die. Linkedin. https://www.linkedin.com/pulse/20130917155206-69244073-say-goodbye -to-mbti-the-fad-that-won-t-die Guntuku, Sharath… ↗
  18. Kiritchenko, Svetlana and Saif Mohammad. 2018. Examining gender and race bias in two hundred sentiment analysis systems. In Proceedings of the Seventh Joint Conference on Lexical and Computational Semantics, pages 43–53. https://doi.org/10.18653/v1/S18-2005 Klenner, Manfred, Anne Göhring, Michael… ↗
  19. Kucher, Kostiantyn, Carita Paradis, and Andreas Kerren. 2018. Visual analysis of sentiment and stance in social media texts. In EuroVis (Posters), pages 49–51. find
  20. Lakoff, George. 2008. Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. University of Chicago Press. find
  21. Lazarus, Richard S. 1991. Progress on a cognitive-motivational-relational theory of emotion. American Psychologist, 46(8):819. https://doi.org/10.1037/0003-066X.46 .8.819 Lindsey, Linda L. 2015. The sociology of gender theoretical perspectives and feminist frameworks. In Gender Roles. Routledge, pag… ↗
  22. Lysaght, Tamra, Hannah Yeefen Lim, Vicki Xafis, and Kee Yuan Ngiam. 2019. AI-assisted decision-making in healthcare. Asian Bioethics Review, 11(3):299–314. https://doi.org/10.1007/s41649-019 -00096-0 MacAvaney, Sean, Anjali Mittu, Glen Coppersmith, Jeff Leintz, and Philip Resnik. 2021. Community-lev… ↗
  23. Portland. Mohammad, Saif. 2012. \#Emotional tweets. In *SEM 2012: The First Joint Conference on Lexical and Computational Semantics–Volume 1: Proceedings of the Main Conference and the Shared Task, and Volume 2: Proceedings of the Sixth International Workshop on Semantic Evaluation (SemEval 2012), pa…
  24. Mohammad, Saif, Felipe Bravo-Marquez, Mohammad Salameh, and Svetlana Kiritchenko. 2018. SemEval-2018 task 1: Affect in tweets. In Proceedings of the 12th International Workshop on Semantic Evaluation, pages 1–17. https://doi.org/10.18653/v1/S18-1001 Mohammad, Saif and Svetlana Kiritchenko. 2018. Un… ↗
  25. Mohammad, Saif M. 2018. Obtaining reliable human ratings of valence, arousal, and dominance for 20,000 English words. In Proceedings of the Annual Conference of the Association for Computational Linguistics (ACL), Melbourne, Australia. https://doi.org/10.18653/v1/P18-1017 Mohammad, Saif M. 2020. Pr… ↗
  26. Mohammad, Saif M. 2021a. Ethics sheets for AI tasks. In Proceedings of the 60th Annual Meeting of the Association of Computational Linguistics (ACL-2022). find
  27. Mohammad, Saif M. 2021b. Sentiment analysis: Automatically detecting valence, emotions, and other affectual states from text. In Herbert L. Meiselman, editor, Emotion Measurement (Second Edition), Woodhead Publishing, pages 323–379. https://doi.org/10.1016/B978-0-12 -821124-3.00011-9 Mohammad, Saif… — [doi:10.1016/b978-0-12](https://doi.org/10.1016/b978-0-12) ↗
  28. Mohammad, Saif M. and Peter D. Turney. 2013. Crowdsourcing a word-emotion association lexicon. Computational Intelligence, 29(3):436–465. https://doi.org/10.1111/j.1467-8640.2012 .00460.x Monteiro, Mike. 2019. Ruined by Design: How Designers Destroyed the World, and What We Can Do to Fix It. ↗
  29. Mule Design. Motti, Vivian Genaro and Anna Evmenova. 2020. Designing technologies for neurodiverse users: Considerations from research practice. In Human Interaction and Emerging Technologies, pages 268–274. https://doi.org/10.1007/978-3-030 -25629-6 42 Mozafari, Nika, Welf H. Weiger, and Maik Hamme… — [doi:10.1007/978-3-030](https://doi.org/10.1007/978-3-030) ↗
  30. Mulligan, Deirdre K., Daniel Kluttz, and Nitin Kohli. 2019. Shaping our tools: Contestability as a means to promote responsible algorithmic decision making in the professions. Available at SSRN 3311894. Nielsen, Finn Årup. 2011. A new ANEW: Evaluation of a word list for sentiment analysis in microb… find
  31. Noel, Lesley-Ann. 2016. Promoting an emancipatory research paradigm in design education and practice. In Proceedings of DRS2016 International Conference, Vol. 6: Future–Focused Thinking, pages 27–30. https://doi.org/10.21606/drs.2016 .355 Oliver, Michael. 1997. Emancipatory research: Realistic goal… — [doi:10.21606/drs.2016](https://doi.org/10.21606/drs.2016) ↗
  32. Osgood, Charles Egerton, George J. Suci, and Percy H. Tannenbaum. 1957. The Measurement of Meaning. 47. University of Illinois Press. find
  33. Panesar, Arjun. 2019. Machine Learning and AI for Healthcare. Springer. https://doi.org/10.1007/978-1-4842 -3799-1 Pang, Bo, Lillian Lee, and Shivakumar Vaithyanathan. 2002. Thumbs up? Sentiment classification using machine learning techniques. In Proceedings of the 2002 Conference on Empirical Met… — [doi:10.1007/978-1-4842](https://doi.org/10.1007/978-1-4842) ↗
  34. Perez, Caroline Criado. 2019. Invisible Women: Exposing Data Bias in a World Designed for Men. find
  35. Random House. Picard, Rosalind W. 2000. Affective Computing. MIT Press. https://doi.org/10.7551/mitpress/1140.001.0001 Pinker, Steven. 2007. The Stuff of Thought: Language as a Window Into Human Nature. ↗
  36. Penguin. Poliak, Adam, Jason Naradowsky, Aparajita Haldar, Rachel Rudinger, and Benjamin Van Durme. 2018. Hypothesis only baselines in natural language inference. In Proceedings of the Seventh Joint Conference on Lexical and Computational Semantics, pages 180–191. https://doi.org/10 .18653/v1/S18-20… find
  37. Quercia, Daniele, Jonathan Ellis, Licia Capra, and Jon Crowcroft. 2012. Tracking “Gross community happiness” from tweets. In Proceedings of the ACM 2012 Conference on Computer Supported Cooperative Work, CSCW ’12, pages 965–968. https://doi.org/10 .1145/2145204.2145347 Resnik, Philip, William Armstr… find
  38. Russell, James A. 1980. A circumplex model of affect. Journal of Personality and Social Psychology, 39(6):1161. https://doi.org/10.1037/h0077714 Russell, James A. 2003. Core affect and the psychological construction of emotion. Psychological Review, 110(1):145. https://doi.org/10.1037/0033-295X .11… ↗
  39. Appraisal Theory. John Wiley & Sons Ltd. Schwartz, Hansen Andrew, Johannes C. Eichstaedt, Margaret L. Kern, Lukasz Dziurzynski, Richard E. Lucas, Megha Agrawal, Gregory J. Park, Shrinidhi K. Lakshmikanth, Sneha Jha, Martin EP Seligman, Lyle H. Ungar, and Richard E. Lucas. 2013. Characterizing geogra… find
  40. Schwartz, Roy, Jesse Dodge, Noah A. Smith, and Oren Etzioni. 2020. Green AI. Communications of the ACM, 63(12):54–63. https://doi.org/10.1145/3381831 ↗
  41. Seale, Jane, Melanie Nind, Liz Tilley, and Rohhss Chapman. 2015. Negotiating a third space for participatory research with people with learning disabilities: An examination of boundaries and spatial practices. Innovation: The European Journal of Social Science Research, 28(4):483–497. https://doi.or… ↗
  42. Standing, Susan and Craig Standing. 2018. The ethical use of crowdsourcing. Business Ethics: A European Review, 27(1):72–80. https://doi.org/10.1111/beer.12173 — [doi:10.1111/beer.12173](https://doi.org/10.1111/beer.12173) ↗
  43. Stone, Emma and Mark Priestley. 1996. Parasites, pawns and partners: Disability research and the role of non-disabled researchers. British Journal of Sociology, pages 699–716. https://doi.org/10.2307/591081 Strubell, Emma, Ananya Ganesh, and Andrew McCallum. 2020. Energy and policy considerations f… — [doi:10.1177/0261927x09351676](https://doi.org/10.1177/0261927x09351676) ↗
  44. Wakefield, Jane. 2021. AI emotion-detection software tested on Uyghurs. BBC. https://www.bbc.com/news/technology-57101248 Wiebe, Janyce, Theresa Wilson, and Claire Cardie. 2005. Annotating expressions of opinions and emotions in language. Language Resources and Evaluation, 39(2):165–210. https://do… ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/mohammad-2022-ethics-sheet-automatic-emotion-recognition/