Linguistics 575: Societal Impacts of Language Technology

Autumn Quarter, 2023

Course Info

• Lecture: Wednesdays, 3:30-5:50 in ECE 025 and online (Zoom link in Canvas)
• Course Canvas (discussion board, assignment submission, grades)

Instructor Info

• Emily M. Bender
• Office Hours: (Most) Wednesdays 2-3pm & (most) Fridays 10-11am & by appointment, online (Zoom link in course Canvas)

Syllabus

Description

The goal of this course is to better understand the ethical considerations that arise in the deployment of NLP technology, including how to identify people likely to be impacted by the use of the technology (direct and indirect stakeholders), what kinds of risks the technology poses, and how to design systems in ways that better support stakeholder values.

Through discussions of readings in the growing research literature on fairness, accountability, transparency and ethics (FATE) in NLP and allied fields, and value sensitive design, we will seek to answer the following questions:

• What can go wrong, when we use NLP systems, in terms of specific harms to people?
• How can fix/prevent/mitigate those harms?
• What are our responsibilities as NLP researchers and developers in this regard?

Course projects are expected to take the form of a term paper analyzing some particular NLP task or data set in terms of the concepts developed through the quarter and looking forward to how ethical best practices could be developed for that task/data set.

Prerequisites: Graduate standing. The primary audience for this course is expected to be CLMS students, but graduate students in other programs are also welcome.

Accessibility policies

If you have already established accommodations with Disability Resources for Students (DRS), please communicate your approved accommodations to me at your earliest convenience so we can discuss your needs in this course.

If you have not yet established services through DRS, but have a temporary health condition or permanent disability that requires accommodations (conditions include but not limited to; mental health, attention-related, learning, vision, hearing, physical or health impacts), you are welcome to contact DRS at 206-543-8924 or uwdrs@uw.edu or disability.uw.edu. DRS offers resources and coordinates reasonable accommodations for students with disabilities and/or temporary health conditions. Reasonable accommodations are established through an interactive process between you, your instructor(s) and DRS. It is the policy and practice of the University of Washington to create inclusive and accessible learning environments consistent with federal and state law.

Washington state law requires that UW develop a policy for accommodation of student absences or significant hardship due to reasons of faith or conscience, or for organized religious activities. The UW's policy, including more information about how to request an accommodation, is available at Faculty Syllabus Guidelines and Resources. Accommodations must be requested within the first two weeks of this course using the Religious Accommodations Request form available at https://registrar.washington.edu/students/religious-accommodations-request/.

[Note from Emily: The above language is all language suggested by UW and in the immediately preceding paragraph in fact required by UW. I absolutely support the content of both and am struggling with how to contextualize them so they sound less cold. My goal is for this class to be accessible. I'm glad the university has policies that help facilitate that. If there is something you need that doesn't fall under these policies, I hope you will feel comfortable bringing that up with me as well.]

Requirements

• KWLA paper (approx 7 pages) (15)
• Exercise 1: SciComm (10)
• Exercise 2: Ethical Considerations Section (10)
• Participation in discussions (incl. Canvas) (15)
• Peer feedback on term project (5)
• Term project (45)

Schedule of Topics and Assignments (still subject to change)

NOTE: still need to place the ethics statement assignment.

Date	Topic	Reading	Due
9/27	Introduction, organization Why are we here? What do we hope to accomplish?	No reading assumed for first day
9/29			KWLA papers: K & W due 11pm
10/4	Foundational readings	Choose two articles from Foundations below, and be prepared to discuss our reading questions: What "common sense" do we take for granted when thinking about "AI" technology that might not necessarily be true? How does technology exacerbate societal harms? (examples and mechanisms) How do we define "ethical" or "unethical"? What expertise are the authors writing from? How much do people who are benefitting from the tech know about how it's harming others? How do we think about identity categories/types of communities in this space and which ones are particularly salient? How do we identify affected communities?
10/11	Value sensitive design	Choose two articles from Value sensitive design below, and be prepared to discuss our reading questions: What kinds of values recur across different projects? What values are specific to different communities? How does "values" differ from "ethics"? How do you evaluate the success of a design process/product in terms of supporting values? Is value sensitive design intended to reduce harm? If so, how? Who is considered the target demographic, within a given a value sensitive design study? How can these methodologies be applied in computational linguistics? What does the opposite look like? What is value agnostic design?
10/18	Bias and discrimination	Choose two articles from Bias/Discrimination below, and be prepared to discuss our reading questions: How is bias defined and measured? Is it possible to over-interpret measures of bias? What kind of taxonomies of harm arising from bias could we create? How do the papers address intersectional identities? Where is the line between generalizability and discrimination? What are the ways that bias sneaks in? What solutions do the authors offer?
10/25	Language variation and emergent bias Translation technologies	Choose two articles from Language Variation and Emergent Bias and/or Translation Technologies below, and be prepared to discuss our reading questions: How well does model adaptation work, especially across related languages, to boost performance in low-resource languages? How are language attitudes reflected in the design of NLP tools? How does knowledge of sociolinguistics help in the improvement of cross-variety performance? How do MT products affect the study of translation and literature? What kinds of biases can machine translation amplify? How can these be mitigated? What positive and negative impacts does MT have on the world? How can negative impacts be mitigated? What are the risks of using MT as opposed to human translators? Bonus, perhaps not well-matched to these readings: How do tech policies accommodate different languages and their different needs?
10/27			Term paper proposals due
11/1	Scicomm and ethics education	Choose two articles from SciComm and Ethics Education below, and be prepared to discuss our reading questions: How can we account for cultural differences when doing science communication? How might technology (internet, NLP, social media) make science communication easier or more difficult? How do you decide what to communicate about in science communication? (How do you know, "Here's something I need to tell the public about?") How do you decide what level of detail to include? What are the incentives for researchers to engage in good science communication? How do you get the science communication artifact out where it can be seen? What are the pushbacks against ethics education? What's the current state of ethics education in NLP? What is/isn't being talked about? Why are we currently lacking in ethics education in NLP?
11/3			Scicomm exercise due
11/8	Documentation and transparency	Choose two articles from Documentation and Transparency below, and be prepared to discuss our reading questions: What kinds of metadata are requested by the documentation toolkits? How should the information be collected? Who has access to the documentation? How can transparency benefit others in the field & how can it benefit the public? (What makes documentation actionable?) How does the process of documentation help dataset & model creators think about societal impact? How are data documentation toolkits evaluated? What ethical considerations are raised about data and model documentation? How the has practice of dataset and model documentation changed over time? How can documentation be kept current with information needs? What incentives and disincentives exist for doing dataset & model documentation?
11/13			Term paper outline due
11/15	Content moderation and toxicity detection	Choose two articles from Content moderation and toxicity detection below, and be prepared to discuss our reading questions: What metrics are used to classify speech as hateful or toxic? What language varieties are more or less likely to be classified as abusive/hateful? What strategies can be employed to recognize and handle culturally specific forms of hate speech and abusive language? What steps are taken to prevent automated content moderation systems from performing unintended censorship? What areas are often neglected by existing toxicity detections systems and taxonomies of abuse & harm? How do these systems enforce existing systems of power and how can that be prevented? Whose perspective are privileged? How important is model transparency in NLP applications for content moderation, and how can it be achieved?
11/22	Policy, regulation and guidelines, Ethics statements	Choose one article from Changing Practice: Policy, regulation, and guidelines below, and be prepared to discuss our reading questions (as relevant to that piece). Also, bring your draft ethical considerations section. What are the differences between guidelines, guardrails, principles, policies, regulations, and how are they made binding or non-binding and how are they decided on? Who is the target audience of the policy/guidelines/etc.? How have policy proposals evolved/adapted to new technologies over time? How are these policies enforced/what are the incentives or consequences involved? Who is responsible for proposing policies? Whose else is getting their interests/perspectives consulted?	Ethical considerations section draft (bring to class)
11/27			Ethical considerations section
11/29	Privacy	Choose two articles from Privacy below, and be prepared to discuss our reading questions: What is PII and what does it include? What does it leave out? What communities may be disproportionately impacted by privacy concerns? Is privacy an individual or collective right? What personal information isn't really private (practically speaking)? How much is tacit/coerced consent relied on to skirt privacy concerns? What are the issues with this? How to navigate tensions between privacy and public safety? Who resolves this and how? What laws/policies are in place to protect privacy/personal information? How can policies be revised? How do privacy issues impact research design decisions and our responsibility as researchers?
12/1			Term paper draft due
12/6	ChatGPT/synthetic media machines	TBD	Comments on partner's paper draft due
12/8			KWLA papers due
12/12			Final papers due 11pm

Bibliography

NOTE This is still very much a work in progress! I have more papers to add, and some of these need to be recategorized.

Overviews/Calls to Action

• Amblard, M. (2016). Pour un TAL responsable. Traitement Automatique des Langues , 57 (2), 21-45.
• boyd danah. (Sept 13, 2019). Facing the great reckoning head-on. Medium.
• Crawford, K., & Calo, R. (2016). There is a blind spot in AI research. Nature, 538 (7625), 311.
• Executive Office of the President National Science and Technology Council Committee on Technology. (2016). Preparing for the future of artificial intelligence. (See especially p. 30)
• Fort, K., Adda, G., & Cohen, K. B. (2016). Ethique et traitement automatique des langues et de la parole: entre truismes et tabous. Traitement Automatique des Langues, 57 (2), 7-19.
• Grissom II, A. (2019). Thinking about how NLP is used to serve power: Current and future trends. /Presentation at Widening NLP 2019. [Slides] [Video]
• Hovy, D., & Spruit, S. L. (2016, August). The social impact of natural language processing. In Proceedings of the 54th annual meeting of the association for computational linguistics (volume 2: Short papers) (pp. 591-598). Berlin, Germany: Association for Computational Linguistics.
  • Another presentation by Hovy on related material
• Lefeuvre-Halftermeyer, A., Govaere, V., Antoine, J.-Y., Allegre, W., Pouplin, S., Departe, J.-P., et al. (2016). Typologie des risques pour une analyse éthique de l'impact des technologies du TAL. Traitement Automatique des Langues, 57 (2), 47-71.
• Markham, A. (May 18, 2016). OKCupid data release fiasco: It's time to rethink ethics education. Data & Society: Points.
• O'Neil, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy. NY: Crown Publishing Group.
• Rogaway, P. (2015). The moral character of cryptographic work.
• Shneiderman, B. (2016). Opinion: The dangers of faulty, biased, or malicious algorithms requires independent oversight. Proceedings of the National Academy of Sciences, 113 (48), 13538-13540.
• Sourour, B. (Nov 13, 2016). The code I'm still ashamed of. Medium.

Foundations

• Alkhatib, A. 2021. To Live in Their Utopia: Why Algorithmic Systems Create Absurd Outcomes
• Baria, A. and Cross, K. 2021. The brain is a computer is a brain: neuroscience's internal debate and the social significance of the Computational Metaphor
• Birhane, A. 2021. Algorithmic injustice: a relational ethics approach
• Birhane, A. 2021. The Impossibility of Automating Ambiguity. Artificial Life, 27(1):44-61.
• Broussard, M. 2023. More than a Glitch: Confronting Race, Gender, and Ability Bias in Tech, MIT Press.
• Calo et al (eds). 2021. Telling Stories: On Culturally Responsive Artificial Intelligence
• Crenshaw, K. (1989). Demarginalizing the intersection of race and sex: A black feminist critique of antidiscrimination doctrine, feminist theory and antiracist politics. U. Chi. Legal f., 139.
• Crenshaw, K. (1990). Mapping the margins: Intersectionality, identity politics, and violence against women of color. Stan. L. Rev., 43, 1241.
• Dai, J. 2020. THE PARADOX OF SOCIALLY RESPONSIBLE COMPUTING The limits and potential of teaching tech ethics
• Ekstrand, M. et al. 2021. Fairness and Discrimination in Information Access Systems
• Hanna, A. and Park, T.M. 2020. Against Scale: Provocations and Resistances to Scale Thinking
• Hao, K. et al. 2022. AI Colonialism, MIT Technology Review.
• Hoffman, A. L. 2020. Terms of Inclusion: Data, Discourse, Violence
• Lewis, J.E. et al, 2020. Indigenous Protocol and Artificial Intelligence
• Martin, K. Forthcoming. Algorithmic Bias and Corporate Responsibility: How companies hide behind the false veil of the technological imperative
• McQuillan, Dan. 2022. Resisting AI: An Anti-fascist Approach to Artificial Intelligence. Bristol University Press.
• Mohamed, S. et al. 2020. Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence
• Noble, S. U. (2018). Algorithms of oppression: How search engines reinforce racism. NYU Press.
• Ògúnrẹ̀mí, T. et al. 2023. Decolonizing NLP for "Low-resource Languages": Applying Abebe Birhane's Relational Ethics, Global Review of AI Community Ethics, 1(1).
• O'Neil, C. 2017. Weapons of Math Destruction, Penguin Random House.
• Paullada, A. et al. 2021. Data and its (dis)contents: A survey of dataset development and use in machine learning research Patterns 2.
• Pipkin, E. 2021. A City Is a City -- Against the metaphorization of data
• Raji, D. 2020. How our data encodes systematic racism MIT Tech Review.
• Sambasivan, N. et al. 2021. "Everyone wants to do the model work, not the data work": Data Cascades in High-Stakes AI. CHI 2021.
• Selbst, A.D. et al. 2019. Fairness and Abstraction in Sociotechnical Systems. FAT* 19.
• Vallor et al. 2018. Overview of Ethics in Tech Practice
• Vinsel, L. 2021. You're Doing It Wrong: Notes on Criticism and Technology Hype
• Whittaker, M. 2021. From Ethics to Organizing: Getting Serious about AI (Video of a talk)
• Zimmermann, A. and Zevenbergen B. 2019. AI Ethics: Seven Traps

Philosophical Underpinnings

• Bartky, S. L. (2002). "Sympathy and solidarity" and other essays (Vol. 32). Rowman & Littlefield.
• Bryson, J. J. (2015). Artificial intelligence and pro-social behaviour. In C. Misselhorn (Ed.), Collective agency and cooperation in natural and artificial systems: Explanation, implementation and simulation (pp. 281-306). Cham: Springer International Publishing.
• Butler, J. (2005). Giving an account of oneself. Oxford University Press. (Available online, through UW libraries)
• Cho, S., Crenshaw, K. W., & McCall, L. (2013). Toward a field of intersectionality studies: Theory, applications, and praxis. Signs: Journal of Women in Culture and Society, 38 (4), 785-810.
• DeLaTorre, M. A. (2013). Ethics: A liberative approach. Fortress Press. (Available online through UW Libraries; read intro + chapter of choice)
• Edgar, S. L. (2003). Morality and machines: Perspectives on computer ethics. Jones & Bartlett Learning. (Available online through UW libraries)
• Fieser, J., & Dowden, B. (Eds.). (2016). Internet encyclopedia of philosophy: Entries on Ethics
• Liamputtong, P. (2006). Researching the vulnerable: A guide to sensitive research methods. Sage. (Available online, through UW libraries)
• Prabhumoye, S., Mayfield, E., & Black, A. W. (2019, August). Principled frameworks for evaluating ethics in NLP systems. In Proceedings of the 2019 workshop on widening nlp (pp. 118-121). Florence, Italy: Association for Computational Linguistics.
• Quinn, M. J. (2014). Ethics for the information age. Pearson.
• Zalta, E. N. (Ed.). (2019). The Stanford encyclopedia of philosophy (Winter 2016 Edition ed.): Entries on Ethics

Value Sensitive Design and Other Design Approaches

• Borning, A. et al. 2005. Informing Public Deliberation: Value Sensitive Design of Indicators for a Large-Scale Urban Simulation, In: Gellersen, H., Schmidt, K., Beaudouin-Lafon, M., Mackay, W. (eds) ECSCW 2005. Springer, Dordrecht.
• Borning, A., & Muller, M. (2012). Next steps for value sensitive design. In Proceedings of the SIGCHI conference on human factors in computing systems (pp. 1125-1134).
• Friedman, B. (1996). Value-sensitive design. ACM Interactions, 3 (6), 17-23.
• Friedman, B. and Hendry, D. 2012. The envisioning cards: a toolkit for catalyzing humanistic and technical imaginations. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI '12). Association for Computing Machinery, New York, NY, USA, 1145-1148.
• Friedman, B., & Hendry, D. G. (2019). Value sensitive design: Shaping technology with moral imagination. MIT Press.
• Friedman, B., Hendry, D. G., Borning, A., et al. (2017). A survey of value sensitive design methods. Foundations and Trends in Human-Computer Interaction , 11 (2), 63-125.
• Friedman, B., & Kahn Jr., P. H. (2008). Human values, ethics, and design. In J. A. Jacko & A. Sears (Eds.), The human-computer interaction handbook (Revised second ed., pp. 1241-1266). Mahwah, NJ.
• Jacobs, N and Huldtgren, A. 2021. Why value sensitive design needs ethical commitments Ethics and Information Technology 23.
• Leidner, J. L., & Plachouras, V. (2017). Ethical by design: Ethics best practices for natural language processing. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 30-40). Valencia, Spain: Association for Computational Linguistics.
• Miller, J. et al. 2007. Value Tensions in Design: The Value Sensitive Design, Development, and Appropriation of a Corporation's Groupware System, GROUP'07, November 4-7, 2007, Sanibel Island, Florida, USA.
• Nathan, L. P., Klasnja, P. V., & Friedman, B. (2007). Value scenarios: a technique for envisioning systemic effects of new technologies. In CHIགྷ extended abstracts on human factors in computing systems (pp. 2585-2590).
• Schnoebelen, T. (2017). Goal-oriented design for ethical machine learning and NLP. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 88-93). Valencia, Spain: Association for Computational Linguistics.
• Spiekermann, S. & Winkler, T. 2021. Twenty years of value sensitive design: a review of methodological practices in VSD projects Ethics and Information Technology 23.
• Umbrello, S and van de Poel, I. 2021. Mapping value sensitive design onto AI for social good principles AI and Ethics 1.
• Yoo, D. 2021. Stakeholder Tokens: a constructive method for value sensitive design stakeholder analysis Ethics and Information Technology 23.
• Young, M., Magassa, L., & Friedman, B. (2019). Toward inclusive tech policy design: A method for underrepresented voices to strengthen tech policy documents. Ethics and Information Technology, 21(2), 89-103.

Documentation and Transparency

• Arnold, Matthew, Rachel K. E. Bellamy, Michael Hind, Stephanie Houde, Sameep Mehta, Aleksandra Mojsilović, Ravi Nair, Karthikeyan Natesan Ramamurthy, Alexandra Olteanu, David Piorkowski, Darrell Reimer, John Richards, Jason Tsay, and Kush R. Varshney. 2019. FactSheets: Increasing trust in AI services through supplier’s declarations of conformity. IBM Journal of Research and Development 63, 4/5 (2019), 6:1-6:13.
• Bender, E. M., & Friedman, B. (2018). Data statements for natural language processing: Toward mitigating system bias and enabling better science. Transactions of the Association for Computational Linguistics, 6, 587-604.
• Bender, E. M., Friedman, B. and McMillan-Major, A. (2021). A Guide for Writing Data Statements for Natural Language Processing.
• Birhane, A., Prabhu, V.U., and Kahembwe, E. (2021). Multimodal datasets: misogyny, pornography, and malignant stereotypes
• Birhane, A., Han, S., Boddeti, V., & Luccioni, S. (2023). Into the LAION’s Den: Investigating Hate in Multimodal Datasets. In Thirty-seventh Conference on Neural Information Processing Systems Datasets and Benchmarks Track.
• Costanza-Chock, Sasha, Raji, Inioluwa Deborah and Buolamwini, Joy. (2022). Who Audits the Auditors? Recommendations from a field scan of the algorithmic auditing ecosystem, FAccT '22: Proceedings of the 2022 ACM Conference on Fairness, Accountability, p.1571–1583
• Couillault, A. et al. 2014. Evaluating Corpora Documentation with regards to the Ethics and Big Data Charter. LREC 2014.
• Crisan, Anamaria, Margaret Drouhard, Jesse Vig, and Nazneen Rajani. 2022. Interactive Model Cards: A Human-Centered Approach to Model Documentation. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency (FAccT '22). Association for Computing Machinery, New York, NY, USA, 427–439.
• Dodge et al. 2021. Documenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus EMNLP 2021.
• Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé III, H., and Crawford, K. (2018). Datasheets for datasets. Proceedings of the 5th Workshop on Fairness, Accountability, and Transparency in Machine Learning.
• Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé III, H., and Crawford K. (2021). Datasheets for datasets. CACM 64(12):86-92.
• Holland, S., Hosny, A., Newman, S., Joseph, J., & Chmielinski, K. (2018). The dataset nutrition label: A framework to drive higher data quality standards.
• Jo, E. S. and Gebru, T. 2020. Lessons from archives: strategies for collecting sociocultural data in machine learning. FAT* ཐ.
• McMillan-Major, Angelina, Emily M. Bender and Batya Friedman. 2023. Data Statements: From Technical Concept to Community Practice, ACM Journal on Responsible Computing.
• McMillan-Major, Angelina, Salomey Osei, Juan Diego Rodriguez, Pawan Sasanka Ammanamanchi, Sebastian Gehrmann, Yacine Jernite. (2021). Reusable Templates and Guides For Documenting Datasets and Models for Natural Language Processing and Generation: A Case Study of the HuggingFace and GEM Data and Model Cards. Proceedings of the 1st Workshop on Natural Language Generation, Evaluation, and Metrics (GEM 2021), pages 121–135, Online. Association for Computational Linguistics.
• Mieskes, M. (2017, April). A quantitative study of data in the NLP community. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 23-29). Valencia, Spain: Association for Computational Linguistics.
• Mitchell, M., Wu, S., Zaldivar, A., Barnes, P., Vasserman, L., Hutchinson, B., et al. (2019). Model cards for model reporting. In Proceedings of the conference on fairness, accountability, and transparency (pp. 220-229). New York, NY, USA: ACM.
• Moss, E. et al. 2021. Assembling Accountability: Algorithmic Impact Assessment for the Public Interest
• Partnership on AI. (2019). ABOUT-ML: Annotation and benchmarking on understanding and transparency of machine learning lifecycles (ABOUT ML).
• Stoyanovich, Julia and Bill Howe. 2019. Nutritional labels for data and models. A Quarterly bulletin of the Computer Society of the IEEE Technical Committee on Data Engineering 42, 3 (2019).
• White House (Biden Administration). (2023). Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. Search specifically for requirements about data and model dcoumentation.

Other Best Practices

• Amodei, D., Olah, C., Steinhardt, J., Christiano, P., Schulman, J., & Mané, D. (2016). Concrete problems in AI safety. CoRR, abs/1606.06565.
• Kulynych, B. et al. 2020. POTs: protective optimization technologies. FAT* 2020.
• Markham, A. (2012). Fabrication as ethical practice: Qualitative inquiry in ambiguous Internet contexts. Information, Communication & Society, 15(3), 334-353.
• Ratto, M. (2011). Critical making: Conceptual and material studies in technology and social life. The Information Society, 27 (4), 252-260.
• Russell, S., Dewey, D., & Tegmark, M. (2015). Research priorities for robust and benefcial artifcial intelligence. AI Magainze.
• Shilton, K., & Anderson, S. (2016). Blended, not bossy: Ethics roles, responsibilities and expertise in design. Interacting with Computers.
• Shilton, K., & Sayles, S. (2016). "We aren't all going to be on the same page about ethics": Ethical practices and challenges in research on digital and social media. In 2016 49th Hawaii international conference on system sciences (HICSS) (pp. 1909-1918).

Bias/Discrimination

• Antoniak and Mimno. 2021. Bad Seeds: Evaluating Lexical Methods for Bias Measurement ACL 2021.
• Bender, Emily M., Timnit Gebru, Angelina McMillan-Major, and Shmargaret Shmitchell. 2021. On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜. In Proceedings of FAccT 2021, pp.610-623.
• Blodgett, S. et al. 2020. Language (Technology) is Power: A Critical Survey of "Bias" in NLP ACL 2020.
• Blodgett, S. et al. 2021. Stereotyping Norwegian Salmon: An Inventory of Pitfalls in Fairness Benchmark Datasets ACL 2021.
• Bolukbasi, T., Chang, K.-W., Zou, J. Y., Saligrama, V., & Kalai, A. T. (2016). Man is to computer programmer as woman is to homemaker? Debiasing word embeddings. In D. D. Lee, M. Sugiyama, U. V. Luxburg, I. Guyon, & R. Garnett (Eds.), Advances in neural information processing systems 29 (pp. 4349-4357). Curran Associates, Inc.
• Caliskan, A., Bryson, J. J., & Narayanan, A. (2017). Semantics derived automatically from language corpora necessarily contain human biases. Science.
• Caliskan, A., Bryson, J., & Narayanan, A. (2016). A story of discrimination and unfairness. (Talk presented at HotPETS 2016)
• Davani, A.M. et al. 2021. Dealing with Disagreements: Looking Beyond the Majority Vote in Subjective Annotations, to appear in TACL.
• Friedman, B., & Nissenbaum, H. (1996). Bias in computer systems. ACM Transactions on Information Systems (TOIS), 14(3), 330-347.
• Garg, N., Schiebinger, L., Jurafsky, D., & Zou, J. (2018). Word embeddings quantify 100 years of gender and ethnic stereotypes. Proceedings of the National Academy of Sciences , 115 (16), E3635-E3644.
• Goldfarb-Tarrant, S., et al. 2021. Intrinsic Bias Metrics Do Not Correlate with Application Bias. ACL 2021.
• Gonen, H., & Goldberg, Y. (2019). Lipstick on a pig: Debiasing methods cover up systematic gender biases in word embeddings but do not remove them. In Proceedings of the 2019 conference of the north American chapter of the association for computational linguistics: Human language technologies, volume 1 (long and short papers) (pp. 609-614). Minneapolis, Minnesota: Association for Computational Linguistics.
• Guynn, J. (Jun 10, 2016). `Three black teenagers' Google search sparks outrage. USA Today.
• Herbelot, A., Redecker, E. von, & Müller, J. (2012). Distributional techniques for philosophical enquiry. In Proceedings of the 6th workshop on language technology for cultural heritage, social sciences, and humanities (pp. 45-54). Avignon, France: Association for Computational Linguistics.
• Hovy, D. and Prabhumoye, S. 2021. Five sources of bias in natural language processing
• Liang, P. et al. 2021. Towards Understanding and Mitigating Social Biases in Language Models Proceedings of the 38 th International Conference on Machine Learning, PMLR 139, 2021.
• Manzini, T., Yao Chong, L., Black, A. W., & Tsvetkov, Y. (2019). Black is to criminal as caucasian is to police: Detecting and removing multiclass bias in word embeddings. In Proceedings of the 2019 conference of the north American chapter of the association for computational linguistics: Human language technologies, volume 1 (long and short papers) (pp. 615-621). Minneapolis, Minnesota: Association for Computational Linguistics.
• Noble, S. U. (2018). Algorithms of oppression: How search engines reinforce racism. NYU Press.
• Rogers, A. 2021. Changing the World by Changing the Data. ACL 2021.
• Rudinger, R., May, C., & Van Durme, B. (2017). Social bias in elicited natural language inferences. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 74-79). Valencia, Spain: Association for Computational Linguistics.
• Schluter, N. (2018). The word analogy testing caveat. In Proceedings of the 2018 conference of the north American chapter of the association for computational linguistics: Human language technologies, volume 2 (short papers) (pp. 242-246). New Orleans, Louisiana: Association for Computational Linguistics.
• Schmidt, B. (2015). Rejecting the gender binary: A vector-space operation. (Blog post, accessed 12/29/16)
• Shen, J.H. et al. 2018. Darling or Babygirl? Investigating Stylistic Bias in Sentiment Analysis FATML 2018.
• Sweeney, L. (May 1, 2013). Discrimination in online ad delivery. Communications of the ACM, 56 (5), 44-54.
• Talat, Z. et al. 2021. Disembodied Machine Learning: On the Illusion of Objectivity in NLP.
• Xu, A. et al. 2021. Detoxifying Language Models Risks Marginalizing Minority Voices. NAACL 2021.

Fairness

• Hardt, M. (Sep 26, 2014). How big data is unfair: Understanding sources of unfairness in data driven decision making. Medium.
• Rao, D. (n.d.). Fairness in machine learning. (slides)
• Zliobaite, I. (2015). On the relation between accuracy and fairness in binary classification. CoRR, abs/1505.05723.

Other resources on bias

• Angwin, J., & Larson, J. (Dec 30, 2016). Bias in criminal risk scores is mathematically inevitable, researchers say. ProPublica.
• boyd, d. (2015). What world are we building? (Everett C Parker Lecture. Washington, DC, October 20)
• Brennan, M. (2015). Can computers be racist? big data, inequality, and discrimination. (online; Ford Foundation)
• Clark, J. (Jun 23, 2016). Artificial intelligence has a `sea of dudes' problem. Bloomberg Technology.
• Crawford, K. (Apr 1, 2013). The hidden biases in big data. Harvard Business Review.
• Daumé III, H. (Nov 8, 2016). Bias in ML, and teaching AI. (Blog post, accessed
• Larson, J., Angwin, J., & Parris Jr., T. (Oct 19, 2016). Breaking the black box: How machines learn to be racist. ProPublica. 1/17/17)
• Emspak, J. (Dec 29, 2016). How a machine learns prejudice: Artificial intelligence picks up bias from human creators--not from hard, cold logic. Scientific American.
• Jacob. (May 8, 2016). Deep learning racial bias: The avenue Q theory of ubiquitous racism. Medium.

More papers in the Proceedings of the First Workshop on Gender Bias in Natural Language Processing

Demographic variables

• Ahmen, A. 2018. Trans Competent Interaction Design: A Qualitative Study on Voice, Identity, and Technology. Interacting with Computers
• Field, A. et al. 2021. A Survey of Race, Racism, and Anti-Racism in NLP ACL 2021.
• Larson, B. (2017). Gender as a variable in natural-language processing: Ethical considerations. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 1-11). Valencia, Spain: Association for Computational Linguistics.
• Morgan Klaus Scheuerman, Katta Spiel, Oliver L. Haimson, Foad Hamidi, Stacy M. Branham. 2020. HCI Guidelines for Gender Equity and Inclusivity

Chatbots

• Abercrombie, G. et al. 2021. Alexa, Google, Siri: What are Your Pronouns? Gender and Anthropomorphism in the Design and Perception of Conversational Assistants Proceedings of the 3rd Workshop on Gender Bias in Natural Language Processing.
• Cercas Curry, A., & Rieser, V. (2018). #MeToo Alexa: How conversational systems respond to sexual harassment. In Proceedings of the second ACL workshop on ethics in natural language processing (pp. 7-14). New Orleans, Louisiana, USA: Association for Computational Linguistics.
• Dinan, E. et al. 2021. Anticipating Safety Issues in E2E Conversational AI: Framework and Tooling
• Elder, A. (2019). Conversation from beyond the grave? A neo-Confucian ethics of chatbots of the dead. Journal of Applied Philosophy.
• Fessler, Leah. (Feb 22, 2017). SIRI, DEFINE PATRIARCHY: We tested bots like Siri and Alexa to see who would stand up to sexual harassment. Quartz.
• Fung, P. (Dec 3, 2015). Can robots slay sexism? World Economic Forum.
• Mott, N. (Jun 8, 2016). Why you should think twice before spilling your guts to a chatbot. Passcode.
• Lee, N. et al. 2019. Exploring Social Bias in Chatbots using Stereotype Knowledge. Proceedings of the 2019 Workshop on Widening NLP.
• Paolino, J. (Jan 4, 2017). Google home vs Alexa: Two simple user experience design gestures that delighted a female user. Medium.
• Seaman Cook, J. (Apr 8, 2016). From Siri to sexbots: Female AI reinforces a toxic desire for passive, agreeable and easily dominated women. Salon.
• Twitter. (Apr 7, 2016). Automation rules and best practices. (Web page, accessed 12/29/16)
• Yao, M. (n.d.). Can bots manipulate public opinion? (Web page, accessed 12/29/16)

Privacy

• Resources from Tech Policy Press on privacy and the European GDPR
• Burgess, M. 2023. ChatGPT Has a Big Privacy Problem WIRED.
• Robertson, A. 2023. ChatGPT returns to Italy after ban, VentureBeat.
• UK Information Commissioner's Office (2023) Guidance on Privacy-enhancing technologies (PETs)
• ACLU of Illinois. nd. Episode 16: Protecting Your Biometric Information.

• Abadi, M., Chu, A., Goodfellow, I., Brendan McMahan, H., Mironov, I., Talwar, K., et al. (2016). Deep Learning with Differential Privacy. ArXiv e-prints.
• Amazon.com. 2017. Memorandum of Law in Support of Amazon's Motion to Quash Search Warrant
• Asher-Schapiro, A. and Sherfinski, D. 2021. INSIGHT-AI surveillance takes U.S. prisons by storm Thomson Reuters Foundation News. November 16, 2021.
• Brant, T. (Dec 27, 2016). Amazon Alexa data wanted in murder investigation. PC Mag.
• Carlini, N. et al. 2020. Extracting Training Data from Large Language Models.
• Friedman, B., Kahn Jr, P. H., Hagman, J., Severson, R. L., & Gill, B. (2006). The watcher and the watched: Social judgments about privacy in a public place. Human-Computer Interaction, 21(2), 235-272.
• Golbeck, J., & Mauriello, M. L. (2016). User perception of facebook app data access: A comparison of methods and privacy concerns. Future Internet, 8(2), 9.
• Grissom II, A. (2019). Thinking about how NLP is used to serve power: Current and future trends. /Presentation at Widening NLP 2019. [Slides] [Video]
• Lewis, D., Moorkens, J., & Fatema, K. (2017). Integrating the management of personal data protection and open science with research ethics. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 60-65). Valencia, Spain: Association for Computational Linguistics.
• Narayanan, A., & Shmatikov, V. (2010). Myths and fallacies of "personally identifiable information". Communications of the ACM, 53 (6), 24-26.
• Nissenbaum, H. (2009). Privacy in context: Technology, policy, and the integrity of social life. Stanford: Stanford University Press.
• Shilton, K. et al. 2021. Excavating awareness and power in data science: A manifesto for trustworthy pervasive data research. Big Data & Society.
• Solove, D. J. (2007). 'I've got nothing to hide' and other misunderstandings of privacy. San Diego Law Review, 44 (4), 745-772.
• Steel, E., & Angwin, J. (Aug 4, 2010). On the Web's cutting edge, anonymity in name only. The Wall Street Journal.
• Tene, O., & Polonetsky, J. (2012). Big data for all: Privacy and user control in the age of analytics. Northwestern Journal of Technology and Intellectual Property, 11(45), 239-273.
• Vitak, J., Shilton, K., & Ashktorab, Z. (2016). Beyond the Belmont principles: Ethical challenges, practices, and beliefs in the online data research community. In Proceedings of the 19th ACM conference on computer-supported cooperative work & social computing (pp. 941-953).

Social Media

• Association of Internet Researchers. 2019. Internet Research: Ethical Guidelines 3.0
• Dym, B. and Fiesler, C. 2020. Ethical and privacy considerations for research using online fandom data
• Fiesler, C. and Proferes, N. 2018. "Participant" Perceptions of Twitter Research Ethics
• Hallinan, B., Brubaker, J. R., & Fiesler, C. (2019). Unexpected expectations: Public reaction to the facebook emotional contagion study. New Media & Society, 1-19. [Tweet thread]
• Metcalf, J., & Crawford, K. (2016). Where are human subjects in big data research? The emerging ethics divide.Big Data & Society 3(1).
• Roose, K. 2021. Inside Facebook's Data Wars. The New York Times July 14, 2021.
• Shilton, K., & Sayles, S. (2016). ``We aren't all going to be on the same page about ethics'': Ethical practices and challenges in research on digital and social media. In 2016 49th Hawaii international conference on system sciences (HICSS) (pp. 1909-1918).
• Townsend, L., & Wallace, C. (2015). Social media research: A guide to ethics. The University of Aberdeen.
• Tufekci, Z. 2014. Big Questions for Social Media Big Data: Representativeness, Validity and Other Methodological Pitfalls Eighth International AAAI Conference on Weblogs and Social Media.
• Williams, M. L., Burnap, P., & Sloan, L. (2017). Towards an ethical framework for publishing twitter data in social research: Taking into account users views, online context and algorithmic estimation. Sociology, 51 (6), 1149-1168.
• Woodfield, K. (Ed.). (2017). The ethics of online research. Emerald Publishing Limited.
  • See especially Chs 2, 5, 7 and 8

Content moderation/Toxicity detection

• Abercrombie, Gavin, Aiqi Jiang, Poppy Gerrard-abbott, Ioannis Konstas, and Verena Rieser. 2023. Resources for Automated Identification of Online Gender-Based Violence: A Systematic Review. In The 7th Workshop on Online Abuse and Harms (WOAH), pages 170–186, Toronto, Canada. Association for Computational Linguistics.
• Diaz, Mark, Razvan Amironesei, Laura Weidinger, and Iason Gabriel. 2022. Accounting for Offensive Speech as a Practice of Resistance. In Proceedings of the Sixth Workshop on Online Abuse and Harms (WOAH), pages 192–202, Seattle, Washington (Hybrid). Association for Computational Linguistics.
• Dwoskin et al, 2019 Content moderators at YouTube, Facebook and Twitter see the worst of the web — and suffer silently, Washington Post July 25, 2019.
• Jiang, J.A., Scheuerman, M.K., Fiesler, C. and Brubaker, J.R. 2021 Understanding international perceptions of the severity of harmful content online
• Jones, L. 2020. Twitter wants you to know that you're still SOL if you get a death threat -- unless you're President Donald Trump
• Gehman, S. et al. 2020. RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models. Findings of EMNLP.
• Hao, K. and Seetharaman, D. 2023. Cleaning Up ChatGPT Takes Heavy Toll on Human Workers Wall Street Journal. (Podcasts embedded.)
• Kirk, Hannah, Wenjie Yin, Bertie Vidgen, and Paul Röttger. 2023. SemEval-2023 Task 10: Explainable Detection of Online Sexism. In Proceedings of the 17th International Workshop on Semantic Evaluation (SemEval-2023), pages 2193–2210, Toronto, Canada. Association for Computational Linguistics.
• Marshall, B. 2021. Algorithmic misogynoir in content moderation practice
• Perrigo, B. 2023. Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic, TIME.
• Röttger, P. et al. 2021. HateCheck: Functional Tests for Hate Speech Detection Models. ACL 2021.
• Saitov, K. and Derczynski, L. 2021. Abusive Language Recognition in Russian. Proceedings of the 8th Workshop on Balto-Slavic Natural Language Processing.
• Sap, M. et al. 2019. The Risk of Racial Bias in Hate Speech Detection ACL 2019.
• Singh, S., Haridasan, A., & Mooney, R. (2023). “Female Astronaut: Because sandwiches won’t make themselves up there”: Towards Multimodal misogyny detection in memes. In The 7th Workshop on Online Abuse and Harms (WOAH) (pp. 150-159).
• Sigurbergsson, G.I. and Derczynski, L. 2020. Offensive Language and Hate Speech Detection for Danish. Proceedings of the 12th Language Resources and Evaluation Conference.
• Simonite, Tom. 2021. Facebook Is Everywhere; Its Moderation Is Nowhere Close. WIRED
• Talat, Z. et al. 2017. Understanding Abuse: A Typology of Abusive Language Detection Subtasks. ACL 2017.
• Yoder, Michael, Chloe Perry, David Brown, Kathleen Carley, and Meredith Pruden. 2023. Identity Construction in a Misogynist Incels Forum. In The 7th Workshop on Online Abuse and Harms (WOAH), pages 1–13, Toronto, Canada. Association for Computational Linguistics.
• Zampieri, Marcos, Shervin Malmasi, Preslav Nakov, Sara Rosenthal, Noura Farra, and Ritesh Kumar. 2019. Predicting the Type and Target of Offensive Posts in Social Media. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 1415–1420, Minneapolis, Minnesota. Association for Computational Linguistics.
• Zeinert, P., Inie, N. and Derczynski, L. 2021. Annotating Online Misogyny. ACL 2021.

See also the papers in the Proceedings Workshop on Online Abuse and Harms: WOAH 2021, WOAH 2022, WOAH 2023

Crowdsourcing and labor conditions

• Bederson, B. B., & Quinn, A. J. (2011). Web workers unite! Addressing challenges of online laborers. In CHIཇ extended abstracts on human factors in computing systems (pp. 97-106).
• Callison-Burch, C. (2016). Crowd workers. (Slides from Crowdsoucing and Human Computation, accessed online 12/30/16)
• Callison-Burch, C. (2016). Ethics of crowdsourcing. (Slides from Crowdsoucing and Human Computation, accessed online 12/30/16)
• Fort, K., Adda, G., & Cohen, K. B. (2011). Amazon mechanical turk: Gold mine or coal mine? Computational Linguistics, 37 (2), 413-420.
• Irani, L. and Silberman, M.S. 2013. Turkopticon: interrupting worker invisibility in Amazon Mechanical Turk CHI 2013.
• Kummerfeld, J.K. 2021. Quantifying and Avoiding Unfair Qualification Labour in Crowdsourcing. ACL 2021.
• Lefeuvre-Halftermeyer, A.; Govaere, V.; Antoine, J.-Y.; Allegre, W.; Pouplin, S.; Departe, J.-P.; Slimani, S. & Spagnulo, A. Typologie des risques pour une analyse éthique de l'impact des technologies du TAL Revue TAL, ATALA (Association pour le Traitement Automatique des Langues), 2016, 57, 47-71
• Shmueli et al. 2021. Beyond Fair Pay: Ethical Implications of NLP Crowdsourcing. NAACL 2021.
• Snyder, J. (2010). Exploitation and sweatshop labor: Perspectives and issues. Business Ethics Quarterly, 20 (2), 187-213.

Language Variation and Emergent Bias

• Bender, E.M. 2019. The #BenderRule: On Naming the Languages We Study and Why It Matters [audio paper] [works cited]
• Broussard, M. (10 May 2018). Agenda: Why the scots are such a struggle for Alexa and Siri. The Herald.
• Garimella, A., Banea, C., Hovy, D., & Mihalcea, R. (2019). Women's syntactic resilience and men's grammatical luck: Gender-bias in part-of-speech tagging and dependency parsing. In Proceedings of the 57th annual meeting of the association for computational linguistics (pp. 3493-3498). Florence, Italy: Association for Computational Linguistics.
• Mengqing Guo, Jiali Li, Jishun Zhao, Shucheng Zhu, Ying Liu, and Pengyuan Liu. 2022. 中文自然语言处理多任务中的职业性别偏见测量(Measurement of Occupational Gender Bias in Chinese Natural Language Processing Tasks). In Proceedings of the 21st Chinese National Conference on Computational Linguistics, pages 510–522, Nanchang, China. Chinese Information Processing Society of China.
• Hovy, D., & Søgaard, A. (2015). Tagging performance correlates with author age. In Proceedings of the 53rd annual meeting of the Association for Computational Linguistics and the 7th international joint conference on natural language processing (volume 2: Short papers) (pp. 483-488). Beijing, China: Association for Computational Linguistics.
• Huang, X., & Paul, M. J. (2019). Neural user factor adaptation for text classification: Learning to generalize across author demographics. In Proceedings of the eighth joint conference on lexical and computational semantics (*SEM 2019) (pp. 136-146). Minneapolis, Minnesota: Association for Computational Linguistics.
• Jørgensen, A., Hovy, D., & Søgaard, A. (2015). Challenges of studying and processing dialects in social media. In Proceedings of the workshop on noisy user-generated text (pp. 9-18). Beijing, China: Association for Computational Linguistics.
• Joshi, P. et al. 2020. The State and Fate of Linguistic Diversity and Inclusion in the NLP World ACL 2020.
• Jurgens, D., Tsvetkov, Y., & Jurafsky, D. (2017). Incorporating dialectal variability for socially equitable language identification. In Proceedings of the 55th annual meeting of the association for computational linguistics (volume 2: Short papers) (pp. 51-57). Vancouver, Canada: Association for Computational Linguistics.
• Koenecke, A. et al. 2020. Racial disparities in automated speech recognition
• Markl and Lai. 2021. Context-sensitive evaluation of automatic speech recognition: considering user experience & language variation
• Tatman, R. (2017). Gender and dialect bias in YouTube's automatic captions. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 53-59). Valencia, Spain: Association for Computational Linguistics.
• Wassink, A., Gansen, C., and Bartholomew, I. 2022. Uneven success: Automatic speech recognition and ethnicity-related dialects. Speech Communication 140:50-70. (Video of related talk; [slides])

Translation Technologies

• Bar-Hillel, Y. (1960). A Demonstration of the Nonfeasibility of Fully Automatic High Quality Translation (Appendix III). In The present status of automatic translation of languages, (Vol. 1, pp. 158-163).
• Bawden, R. (2018). Going beyond the sentence: Contextual machine translation of dialogue (Doctoral dissertation, Université Paris-Saclay (ComUE)).
• Boitet, C. (1988). PROs and CONs of the pivot and transfer approaches in multilingual Machine Translation. Readings in machine translation, 273-279.
• Braun, S. (2019). Technology and interpreting. Routledge Handbook of Translation and Technology, Routledge, London, 271-288.
• Hatim, B., & Munday, J. (2004). Translation an advanced resource book Routledge.
• Hutchins, J. (2003). ALPAC: the (in) famous report. Readings in machine translation, 14, 131-135.
• Kenny, Dorothy, Joss Moorkens, and Félix Do Carmo. Fair MT: Towards ethical, sustainable machine translation. Translation Spaces 9.1 (2020): 1-11.
• Lehman-Wilzig, S. (2017). Babel and babble: Autonomous, algorithmic, simultaneous translation systems in the glocal village -- consequences & paradoxical outcomes. In S. Brunn & R. Kehrein (Eds.), The Changing World Language Map. New York: Springer Publishing.
• Liu, Lydia H., ليديا ﻫ. ليو, James St. André, جيمس سانت أندريه (2018). The Battleground of Translation: Making Equal in a Global Structure of Inequality / الترجمة في المعترك : البحث عن المساواة في سياق اللامساواة العالمي Alif: Journal of Comparative Poetics, 38, 368–387. http://www.jstor.org/stable/26496380.
• Paullada, Amandalynne. 2021. Machine Translation Shifts Power. The Gradient.
• Ramati, I., & Pinchevski, A. (2017). Uniform multilingualism: A media genealogy of Google Translate. New Media & Society, 1461444817726951.
• Savoldi, Beatrice, Marco Gaido, Luisa Bentivogli, Matteo Negri, Marco Turchi. 2021. Gender Bias in Machine Translation. Transactions of the Association for Computational Linguistics 9:845-874.

Biomedical NLP, Mental Health and Social Media

• Andrade, N. N. Gomes de, Pawson, D., Muriello, D., Donahue, L., & Guadagno, J. (2018, Dec 01). Ethics and artificial intelligence: Suicide prevention on Facebook. Philosophy & Technology, 31(4), 669-684.
• Barnett, I., & Torous, J. (2019, 04). Ethics, Transparency, and Public Health at the Intersection of Innovation and Facebook's Suicide Prevention Efforts. Annals of Internal Medicine, 170 (8), 565-566.
• Benton, A., Coppersmith, G., & Dredze, M. (2017). Ethical research protocols for social media health research. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 94-102). Valencia, Spain: Association for Computational Linguistics.
• Elder, A. (2019). Conversation from beyond the grave? A neo-Confucian ethics of chatbots of the dead. Journal of Applied Philosophy.
• Linthicum, K. P., Schafer, K. M., & Ribeiro, J. D. (2019). Machine learning in suicide science: Applications and ethics. Behavioral sciences & the law , 37 (3), 214-222.
• McKernan, L. C., Clayton, E. W., & Walsh, C. G. (2018). Protecting life while preserving liberty: Ethical recommendations for suicide prevention with artificial intelligence. Frontiers in Psychiatry, 9.
• Suster, S., Tulkens, S., & Daelemans, W. (2017, April). A short review of ethical challenges in clinical natural language processing. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 80-87). Valencia, Spain: Association for Computational Linguistics.
• Tucker, R. P., Tackett, M. J., Glickman, D., & Reger, M. A. (2019). Ethical and practical considerations in the use of a predictive model to trigger suicide prevention interventions in healthcare settings. Suicide and Life-Threatening Behavior , 49 (2), 382-392.

NLP Apps Addressing Ethical Issues/NLP for Social Good

• Demszky, D., Garg, N., Voigt, R., Zou, J., Shapiro, J., Gentzkow, M., et al. (2019). Analyzing polarization in social media: Method and application to tweets on 21 mass shootings. In Proceedings of the 2019 conference of the north American chapter of the association for computational linguistics: Human language technologies, volume 1 (long and short papers) (pp. 2970-3005). Minneapolis, Minnesota: Association for Computational Linguistics.
• Fokkens, A. (2016). Reading between the lines. (Slides presented at Language Analysis Portal Launch event, University of Oslo, Sept 2016)
• Jurgens, D., Hemphill, L., & Chandrasekharan, E. (2019). A just and comprehensive strategy for using NLP to address online abuse. In Proceedings of the 57th annual meeting of the association for computational linguistics (pp. 3658-3666). Florence, Italy: Association for Computational Linguistics.
• Lee, N., Bang, Y., Shin, J., & Fung, P. (2019). Understanding the shades of sexism in popular TV series. In Proceedings of the 2019 workshop on widening NLP (pp. 122-125). Florence, Italy: Association for Computational Linguistics.
• Gershgorn, D. (Feb 27, 2017). NOT THERE YET: Alphabet's hate-fighting AI doesn't understand hate yet. Quartz.
• Google.com. (2017). The women missing from the silver screen and the technology used to find them. Blog post, accessed March 1, 2017.
• Greenberg, A. (2016). Inside Google's Internet Justice League and Its AI-Powered War on Trolls. Wired.
• Kellion, L. (Mar 1, 2017) Facebook artificial intelligence spots suicidal users. BBC News.
• Madnani, N., Loukina, A., Davier, A. von, Burstein, J., & Cahill, A. (2017). Building better open-source tools to support fairness in automated scoring. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 41-52). Valencia, Spain: Association for Computational Linguistics.
• Munger, K. (2016). Tweetment effects on the tweeted: Experimentally reducing racist harassment. Political Behavior, 1-21.
• Munger, K. (Nov 17, 2016). This researcher programmed bots to fight racism on twitter. It worked. Washington Post.
• Murgia, M. (Feb 23, 2017). Google launches robo-tool to flag hate speech online. Financial Times.
• The times is partnering with jigsaw to expand comment capabilities. (Sep 20, 2016). The New York Times.
• Qian, J., ElSherief, M., Belding, E., & Wang, W. Y. (2019). Learning to decipher hate symbols. In Proceedings of the 2019 conference of the north American chapter of the association for computational linguistics: Human language technologies, volume 1 (long and short papers) (pp. 3006-3015). Minneapolis, Minnesota: Association for Computational Linguistics.
• Waseem, Z. (2016). Are you a racist or am I seeing things? Annotator influence on hate speech detection on Twitter. In Proceedings of the first workshop on nlp and computational social science (pp. 138-142). Austin, Texas: Association for Computational Linguistics.
• Waseem, Z., & Hovy, D. (2016). Hateful symbols or hateful people? Predictive features for hate speech detection on Twitter. In Proceedings of the naacl student research workshop (pp. 88-93). San Diego, California: Association for Computational Linguistics.
• Wiegand, M., Ruppenhofer, J., & Kleinbauer, T. (2019). Detection of Abusive Language: the Problem of Biased Datasets. In Proceedings of the 2019 conference of the north American chapter of the association for computational linguistics: Human language technologies, volume 1 (long and short papers) (pp. 602-608). Minneapolis, Minnesota: Association for Computational Linguistics.

• Fake News Challenge
• Jigsaw Challenges
• Perspective (from Jigsaw)
  • But see: Hosseini, H, S. Kannan, B. Zhang and R. Poovendran. 2017. Deceiving Google's Perspective API Built for Detecting Toxic Comments. ArXiv.
• Textio See also:
  • CEO Kieran Snyder's posts on medium.com
  • Recording of Kieran Snyder's NLP Meetup talk from Aug 15, 2016

Other Issues in NLP: Carbon Emissions, Generation, ...

• Radford, A., Wu, J., Child, R., Luan, D., Amodei, D., & Sutskever, I. (2019). Language models are unsupervised multitask learners. OpenAI Blog, 1(8).
• Seabrook, J. (2019). The next word: Where will predictive text take us? The New Yorker, 14 Oct 2019.
• Parra Escartín, C., Reijers, W., Lynn, T., Moorkens, J., Way, A., & Liu, C.-H. (2017). Ethical considerations in NLP shared tasks. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 66-73). Valencia, Spain: Association for Computational Linguistics.
• Smiley, C., Schilder, F., Plachouras, V., & Leidner, J. L. (2017). Say the right thing right: Ethics issues in natural language generation systems. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 103-108). Valencia, Spain: Association for Computational Linguistics.
• Strubell, E., Ganesh, A., & McCallum, A. (2019). Energy and policy considerations for deep learning in NLP. In Proceedings of the 57th annual meeting of the association for computational linguistics (pp. 3645-3650). Florence, Italy: Association for Computational Linguistics.
• Šuster, S., Tulkens, S., & Daelemans, W. (2017). A short review of ethical challenges in clinical natural language processing. In Proceedings of the first ACL workshop on ethics in natural language processing (pp. 80-87). Valencia, Spain: Association for Computational Linguistics.
• Torbati, Y. (Sept, 2019). Google says Google Translate can't replace human translators. Immigration officials have used it to vet refugees. ProPublica.
• Vincent, J. (Feb, 2019). AI researchers debate the ethics of sharing potentially harmful programs. The Verge.

SciComm and Ethics Education

• Burns, T. W., O'Connor, D. J., & Stocklmayer, S. M. (2003). Science communication: A contemporary definition. Public Understanding of Science, 12(2), 183-202.
• Dai, J. 2020. THE PARADOX OF SOCIALLY RESPONSIBLE COMPUTING The limits and potential of teaching tech ethics
• Di Bari, M., & Gouthier, D. (2002). Tropes, science and communication. Journal of Communication, 2(1).
• Fiesler, C., Garrett, N., & Beard, N. (2020). What do we teach when we teach tech ethics?Links to an external site. A syllabi analysis. In Proceedings of the 51st ACM technical symposium on computer science education (pp. 289-295).
• Fischhoff, B. (2013). The sciences of science communication. Proceedings of the National Academy of Sciences, 110(Supplement 3), 14033-14039.
• Gero, K.I. et al. (2021). What Makes Tweetorials Tick: How Experts Communicate Complex Topics on Twitter, Proc. ACM Hum.-Comput. Interact. 5, CSCW2, Article 422.
• Mooney, C. (2010). Do scientists understand the public? American Academy of Arts & Sciences.
• Ngumbi, E. (2018, January 26). If you want to explain your science to the public, here's some advice. Scientific American.
• Phillips, C. M. L., & Beddoes, K. (2013). Really changing the conversation: The deficit model and public under-standing of engineering. In Proceedings of the 120th ASEE annual conference & exposition.
• Raji, Inioluwa Deborah, Morgan Klaus Scheuerman, and Razvan Amironesei. 2021. You Can't Sit With Us: Exclusionary Pedagogy in AI Ethics Education. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (FAccT '21). Association for Computing Machinery, New York, NY, USA, 515–525.
• Shepherd, M. (2016, November 22). 9 tips for communicating science to people who are not scientists. Forbes, 1-4.
• Simis, M. J., Madden, H., Cacciatore, M. A., & Yeo, S. K. (2016). The lure of rationality: Why does the deficit model persist in science communication? Public Understanding of Science, 25 (4), 400-414.
• Young, E. 2021. What Even Counts as Science Writing Anymore? The Atlantic.

• ACM Ethics Task Force. (2016). Code 2018 | ACM ethics. (Web page, accessed 1/5/17)
• The IEEE Global Initiative for Ethical Considerations in Artificial Intelligence and Autonomous Systems. (2016). Ethically aligned design: A vision for prioritizing human wellbeing with artificial intelligence and autonomous systems (AI/AS) (Version 1 -- For Public Discussion).
• ACL Ethics FAQ. 2021. EMNLP version
• US FDA. 2021. Good Machine Learning Practice for Medical Device Development: Guiding Principles

• AI Bill of Rights (White House proposal)
• Algorithmic Accountability Act (2022; not passed)
• FTC Chair Khan and Officials from DOJ, CFPB and EEOC Release Joint Statement on AI
• White House (Biden Administration). (2023). Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. Search specifically for requirements about data and model dcoumentation.
• NIST AI Risk Management Framework
• OMB Draft guidance on AI

• Ashurst, C. et al. 2020. A Guide to Writing the NeurIPS Impact Statement
• Ashurt, C. et al. 2022. AI Ethics Statements: Analysis and Lessons Learnt from NeurIPS Broader Impact Statements FAccT 2022.
• Daumé III, H. (Dec 12, 2016). Should the NLP and ML Communities have a Code of Ethics? (Blog post, accessed 12/30/16)
• Do et al. 2023. “That’s important, but...”: How Computer Science Researchers Anticipate Unintended Consequences of Their Research Innovations, CHI 2023.
• Etlinger, S., & Groopman, J. (2015). The trust imperative: A framework for ethical data use.
• Liu et al, 2022. Examining Responsibility and Deliberation in AI Impact Statements and Ethics Reviews, AIES 2022.
• Moran, C. 2021. Machine Learning, Ethics, and Open Source Licensing (Part II/II)
• Nanayakkara, P. et al. 2021. Unpacking the Expressed Consequences of AI Research in Broader Impact Statements
• Patton, D. U. (16 April 2018). Ethical guidelines for social media research with vulnerable groups. Medium.
• Rakova et al. 2021. Where Responsible AI meets Reality: Practitioner Perspectives on Enablers for Shifting Organizational Practices
• Schwartz et al. 2021. A Proposal for Identifying and Managing Bias in Artificial Intelligence
• Stix, C. 2021. Actionable Principles for Artificial Intelligence Policy: Three Pathways

Ethics Statements

Reading notes

• Schmaltz 2018 is a proposal around this practice. The other references here are examples of paper including an ethics statment. For this week, we'll be writing our own ethics statements, either for our own papers or for others we have selected.

Papers:

• Al-khazra ji, S., Berke, L., Kafle, S., Yeung, P., & Huenerfauth, M. (2018). Modeling the speed and timing of american sign language to generate realistic animations. In Proceedings of the 20th international ACM SIGACCESS conference on computers and accessibility (pp. 259-270). New York, NY, USA: ACM.
• Chen, H., Cai, D., Dai, W., Dai, Z., & Ding, Y. (2019). Charge-based prison term prediction with deep gating network. (To appear at EMNLP 2019)
• Schmaltz, A. (2018). On the utility of lay summaries and AI safety disclosures: Toward robust, open research oversight. In Proceedings of the second ACL workshop on ethics in natural language processing (pp. 1-6). New Orleans, Louisiana, USA: Association for Computational Linguistics.

Synthetic Media Machines

• Anthony Cuthbertson, June 2023. ChatGPT ‘grandma exploit’ gives users free keys for Windows 11
• Jason Koebler, Nov. 2023. Google Researchers’ Attack Prompts ChatGPT to Reveal Its Training Data
• Lilian Weng, Oct. 2023. Adversarial Attacks on LLMs
• Johann Rehberger, Nov. 2023. Hacking Google Bard - From Prompt Injection to Data Exfiltration
• Jiawei Zhou, Yixuan Zhang, Qianni Luo, Andrea G Parker, and Munmun De Choudhury. 2023. Synthetic Lies: Understanding AI-Generated Misinformation and Evaluating Algorithmic and Human Solutions.
• What happens when ChatGPT starts to feed on its own writing? Vox, April 10, 2023.
• A Comparative Analysis between Human-in-the-loop Systems and Large Language Models for Pattern Extraction Tasks
• Human-in-the-loop Machine Translation with Large Language Model
• Training language models to follow instructions with human feedback
• Can Artificial Intelligence Reprogram the Newsroom? Trust, Transparency, and Ethics in Automated Journalism
• A Pathway Towards Responsible AI Generated Content
• US Copyright Office. International Copyright Issues and Artificial Intelligence (Webinar)
• Workers are worried about AI taking their jobs. Artists say it's already happening. Business Insider, October 1, 2023
• Shan, Shawn, Jenna Cryan, Emily Wenger, Haitao Zheng, Rana Hanocka, Ben Y. Zhao. 2023. Glaze: Protecting Artists from Style Mimicry by Text-to-Image Models. Proceedings of USENIX Security Symposium.
• Shah, Chirag and Emily M. Bender. Under review. Envisioning Information Access Systems: What Makes for Good Tools and a Healthy Web?
• Kirchenbauer, J., Geiping, J., Wen, Y., Katz, J., Miers, I., & Goldstein, T. (2023). A Watermark for Large Language Models. ICML.
• We tested a new ChatGPT-detector for teachers. It flagged an innocent student. Washington Post. April 1, 2023.
• These lawyers used ChatGPT to save time. They got fired and fined. Washington Post. November 16, 2023
• Authors' lawsuit against OpenAI could 'fundamentally reshape' artificial intelligence, according to experts
• Judge bins AI copyright lawsuit against DeviantArt, Midjourney – Stability still in the mix
• Artists take new shot at Stability, Midjourney in updated copyright lawsuit
• Evidence mounts as new artists jump on Stability AI, Midjourney copyright lawsuit
• Evaluating LLM outputs
• A Survey on Evaluation of Large Language Models
• A Comparative Analysis between Human-in-the-loop Systems and Large Language Models for Pattern Extraction Tasks
• Bashir, D. (2022). AI literacy : Understanding shifts in our digital ecosystem New Degree Press.
• Schüller, K. (2022). Data and AI literacy for everyone. Statistical Journal of the IAOS, 38(2), 477–490.
• Friedman, Jane. 2023. I Would Rather See My Books Get Pirated Than This (Or: Why Goodreads and Amazon Are Becoming Dumpster Fires).

Other Readings

• Basile, V. nd. The Perspectivist Data Manifesto
• Cohen, K. B., Pestian, J., & Fort, K. (2015). Annotateurs volontaires investis et éthique de l'annotation de lettres de suicidés. In ETeRNAL (ethique et traitement automatique des langues).
• Dalzell, S. 2020. Federal Government used Google Translate for COVID-19 messaging aimed at multicultural communities
• Fort, K., & Couillault, A. (2016). Yes, we care! results of the ethics and natural language processing surveys. In Proceedings of the tenth international conference on language resources and evaluation (LREC 2016). Paris, France: European Language Resources Association (ELRA).
• Gillespie, T. (2014). The relevance of algorithms. In T. Gillespie, P. J. Boczkowski, & K. A. Foot (Eds.), Media technologies: Essays on communication, materiality, and society (pp. 167-194). MIT Press.
• Hardt, M., Price, E., & Srebro, N. (2016). Equality of opportunity in supervised learning. (Accessed online, 12/30/16)
  • Related blog post
• Kleinberg, J. M., Mullainathan, S., & Raghavan, M. (2016). Inherent trade-offs in the fair determination of risk scores. CoRR, abs/1609.05807.
• Metcalf, J., Keller, E. F., & boyd, d. (2016). Perspectives on big data, ethics, and society. (Accessed 12/30/16)
• Meyer, M. N. (2015). Two cheers for corporate experimentation: The A/B illusion and the virtues of data-driven innovation. Colo. Tech. L.J., 13, 273.
• Sloane et al. 2021. A Silicon Valley Love Triangle: Hiring Algorithms, Pseudo- Science, and the Quest for Auditability
• Wallach, H. (Dec 19, 2014). Big data, machine learning, and the social sciences: Fairness, accountability, and transparency. Medium.
• Wattenberg, M., Viégas, F., & Hardt, M. (Oct 7, 2016). Attacking discrimination with smarter machine learning.

Links

Conferences/Workshops

• ACM FAT* Conference 2020 Barcelona, Spain, January 2020
• ALW3: 3rd Workshop on Abusive Language Online at ACL 2019, Florence, Italy, August 2019
• 1st ACL Workshop on Gender Bias for Natural Language Processing at ACL 2019, Florence, Italy, August 2019
• ACM FAT* Conference 2019 Atlanta, GA, January 2019
• ALW2: 2nd Workshop on Abusive Language Online at EMNLP 2018, Brussels, Belgium, October 2018
• FAT/ML 2018 at ICML 2018, Stockholm, Sweden, July 2018
• Ethics in Natural Language Processing at NAACL 2018, New Orleans LA, June 2018
• ACM FAT* Conference 2018 NY, NY, February 2018
• FAT/ML 2017 at KDD 2017, Halifax, Canada, August 2017
• ALW1: 1st Workshop on Abusive Language Online at ACL 2017, Vancouver, Canada, August 2017
• Ethics in Natural Language Processing at EACL 2017, Valencia, Spain, April 2017
• 3rd International Workshop on AI, Ethics and Society 4th or 5th February 2017 San Francisco, USA
• PDDM16 The 1st IEEE ICDM International Workshop on Privacy and Discrimination in Data Mining December 12, 2016 - Barcelona
• Machine Learning and the Law NIPS Symposium 8 December, 2016 Barcelona, Spain
• FAT/ML 2016 NY, NY, November 2016
• AAAI Fall Symposium on Privacy and Language Technologies, November 2016
• Workshop on Data and Algorithmic Transparency (DATཌ) November 19, 2016, New York University Law School
• WSDM 2016 Workshop on the Ethics of Online Experimentation, February 22, 2016 San Francisco, California
• ETHI-CA2 2016: ETHics In Corpus Collection, Annotation and Application LREC 2016, Protoroz, Slovenia.
• FAT/ML 2015, at ICML 2015, Lille, France, July 2015
• ETeRNAL - Ethique et TRaitemeNt Automatique des Langues June 22, 2015, Caen
• Éthique et Traitement Automatique des Langues, Journée d'étude de l'ATALA Paris, France, November 2014
• Fairness, Accountability, and Transparency in Machine Learning (FAT/ML) 2014, at NeurIPS 2014, Montréal, Canada, December 2014

Other lists of resources

• Critical Algorithm Stuides
• FATML resources page
• The Responsible Conduct of Computational Modeling and Research NSF funded project

Other courses

• ACL wiki page on teaching Ethics in NLP (list of courses)
• A Course on Fairness, Accountability and Transparency in Machine Learning (Suresh Venkatasubramanian)
• Ethics for the Information Age (Michael Quinn)
• Another list of courses from Thomast Morgan Jr. at ETSU
• The Dark Side of NLP: Gefahren automatischer Sprachverarbeitung (Michael Strube)
• Histories of AI from Jonnie Penn, Rebecca Charbonneau, Jenny Carla Moran at the University of Cambridge.