Like Trainer, Like Bot? Inheritance of Bias in Algorithmic Content Moderation
Open accessThe DOI matches its Crossref record; 7 of the 7 references it cites were found in a registry or an index.
Abstract
The internet has become a central medium through which `networked publics' express their opinions and engage in debate. Offensive comments and personal attacks can inhibit participation in these spaces. Automated content moderation aims to overcome this problem using machine learning classifiers trained on large corpora of texts manually annotated for offence. While such systems could help encourage more civil debate, they must navigate inherently normatively contestable boundaries, and are subject to the idiosyncratic norms of the human raters who provide the training data. An important objective for platforms implementing such measures might be to ensure that they are not unduly biased towards or against particular norms of offence. This paper provides some exploratory methods by which the normative biases of algorithmic content moderation systems can be measured, by way of a case study using an existing dataset of comments labelled for offence. We train classifiers on comments labelled by different demographic subsets (men and women) to understand how differences in conceptions of offence between these groups might affect the performance of the resulting models on various test sets. We conclude by discussing some of the ethical choices facing the implementers of algorithmic moderation systems, given various desired levels of diversity of viewpoints amongst discussion participants.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Answering the Call for a Standard Reliability Measure for Coding Data
Hayes, A. F., & Krippendorff, K. (2007) · Communication Methods and Measures · doi:10.1080/19312450709336664 · Cited by 4,225 · cited 1×
“Considering the dataset before training classifiers, we measured agreement on toxicity within male and female annotators with Krippendorff’s alpha (due to amenability to missing data, see [20])) 3 , and average measures of how toxicity scores per comment differs between male and female annotators.”
MethodTokunaga, R. S. (2010) · Computers in Human Behavior · doi:10.1016/j.chb.2009.11.014 · Cited by 2,566
Gillespie, T. (2010) · New Media & Society · doi:10.1177/1461444809342738 · Cited by 2,522 · cited 1×
“Online platforms, as ‘curators of public discourse’ [18] or digital extensions of the public sphere [12], have become important spaces for opinion and debate.”
BackgroundCertifying and Removing Disparate Impact
Feldman, M., Friedler, S. A., Moeller, J., Scheidegger, C., & Venkatasubramanian, S. (2015) · Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining · doi:10.1145/2783258.2783311 · Cited by 1,827
Ex Machina: Personal Attacks Seen at Scale
Wulczyn, E., Thain, N., & Dixon, L. (2017) · Proceedings of the 26th International Conference on World Wide Web · doi:10.1145/3038912.3052591 · Cited by 636
What is a flag for? Social media reporting tools and the vocabulary of complaint
Crawford, K., & Gillespie, T. (2016) · New Media & Society · doi:10.1177/1461444814543163 · Cited by 496
Why Unbiased Computational Processes Can Lead to Discriminative Decision Procedures
Calders, T., & Žliobaitė, I. (2013) · Studies in Applied Philosophy, Epistemology and Rational Ethics · doi:10.1007/978-3-642-30487-3_3 · Cited by 128 · cited 1×
“In these fields, efforts to understand and mitigate illegal bias or general unfairness in areas from loan acceptance to word embeddings have propagated [4, 7, 8, 16].”
Background
Cited by
Content moderation, AI, and the question of scale
Gillespie, T. (2020) · Big Data & Society · doi:10.1177/2053951720943234 · Cited by 437
Directions in abusive language training data, a systematic review: Garbage in, garbage out
Vidgen, B., & Derczynski, L. (2020) · PLOS ONE · doi:10.1371/journal.pone.0243300 · Cited by 318
Administration by Algorithm? Public Management meets Public Sector Machine Learning
Veale, M., & Brass, I. (2019) · doi:10.31235/osf.io/mwhnb · Cited by 316
Self-citationHuman-Machine Collaboration for Content Regulation: The Case of Reddit Automoderator
Jhaver, S., Birman, I., Gilbert, E., & Bruckman, A. (2019) · ACM Transactions on Computer-Human Interaction · doi:10.1145/3338243 · Cited by 289
Problematic Machine Behavior: A Systematic Literature Review of Algorithm Audits
Bandy, J. (2021) · Proceedings of the ACM on Human-Computer Interaction · doi:10.1145/3449148 · Cited by 193
Metadata
- Authors
- Binns, R., Veale, M., Van Kleek, M., & Shadbolt, N.
- Venue
- Lecture Notes in Computer Science
- Publisher
- Springer International Publishing
- Published
- 2017
- Pages
- 405-415
- ISSN
- 0302-9743 (ISSN-L)
- 1611-3349
- OpenAlex
- W2730390070
- Cited by
- 81
- References
- 7
- FWCI
- 1.52
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Like Trainer, Like Bot? Inheritance of Bias in Algorithmic Content Moderation. https://app.cochat.ai/papers/10.1007/978-3-319-67256-4_32
- The paper
- Binns, R., Veale, M., Van Kleek, M., & Shadbolt, N. (2017). Like Trainer, Like Bot? Inheritance of Bias in Algorithmic Content Moderation. Lecture Notes in Computer Science. https://doi.org/10.1007/978-3-319-67256-4_32
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.