pith. sign in

arxiv: 2605.22435 · v1 · pith:J4IHSYDUnew · submitted 2026-05-21 · 💻 cs.CL

Assisted Counterspeech Writing at the Crossroads of Hate Speech and Misinformation

Genoveffa Martone , Helena Bonaldi , Marco Guerini This is my paper

Pith reviewed 2026-05-22 07:02 UTC · model grok-4.3

classification 💻 cs.CL
keywords counterspeech generationhate speechmisinformationlarge language modelsfact-checkingexpert revisioncrowdsourced evaluationdataset
0
0 comments X

The pith

A mixed strategy drawing on both fact-checkers and NGOs produces the most effective counterspeech when hate speech and misinformation appear together.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper investigates how large language models can help experts write counterspeech in online settings where hate speech and misinformation reinforce each other. It compares three ways of supplying knowledge to the models: fact-checking articles plus guidelines, NGO reports plus guidelines, and a combination of both. After experts revise the model outputs, crowdsourced assessments find that the combined approach yields the strongest results by correcting facts while also reducing stereotypes and maintaining an empathetic tone. This matters because separate treatments of hate and misinformation have left a gap in handling their frequent overlap at scale. The authors also release a dataset of such claims paired with verified counterspeech for further use.

Core claim

When hate speech and misinformation co-occur, large language models prompted with a mixed set of fact-checkers' guidelines and articles together with NGOs' guidelines and reports generate counterspeech that, after expert revision, scores highest in crowdsourcing evaluations for factual correction, stereotype mitigation, and empathetic engagement.

What carries the argument

The mixed knowledge-driven generation strategy that combines guidelines and source documents from both fact-checkers and NGOs to prompt the LLM.

If this is right

  • Expert post-editing raises the quality of LLM-generated counterspeech in naturalness, exhaustiveness, and guideline adherence even when the initial output is only adequate about 40 percent of the time.
  • The mixed strategy succeeds specifically by delivering factual corrections alongside stereotype mitigation and empathetic engagement.
  • The released dataset of hateful and misinformed claims with expert-verified counterspeech supports additional research on combined phenomena.
  • Human and automatic metrics together confirm that single-source strategies underperform the mixed approach in the evaluated crowdsourcing setting.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same mixed-knowledge prompting pattern could be tested on other overlapping harms such as conspiracy theories paired with targeted harassment.
  • Real-time deployment would require checking how well the approach scales when the underlying fact-checks and NGO reports are updated daily.
  • The dataset could serve as training material for models that learn to produce counterspeech without explicit external knowledge at inference time.

Load-bearing premise

The selected fact-checking articles and NGO reports supply representative, unbiased, and sufficient knowledge to guide high-quality counterspeech generation across diverse cases of co-occurring hate and misinformation.

What would settle it

A follow-up test applying the three strategies to a fresh set of hateful and misinformed claims outside the original knowledge sources, then checking whether crowdsourced raters still rate the mixed version highest on factual accuracy, stereotype reduction, and empathy.

Figures

Figures reproduced from arXiv: 2605.22435 by Genoveffa Martone, Helena Bonaldi, Marco Guerini.

Figure 1
Figure 1. Figure 1: An example of hateful and misinformed claim with CS obtained with the three tested strate￾gies. The text highlighted in orange represents the misinformation-related text, while the light blue refers to the hateful-related text. hateful and misinformed statements targeting six marginalized groups, as well as generated and ex￾pert post-edited versions of the CS and support￾ing external knowledge, are availab… view at source ↗
Figure 2
Figure 2. Figure 2: The pipeline of this work: first, the external knowledge and guidelines are collected. Then, they are used [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Guidelines for the annotation of the CS generated with the mixed configuration. [PITH_FULL_IMAGE:figures/full_fig_p015_3.png] view at source ↗
read the original abstract

Hate speech and misinformation frequently co-occur online, amplifying prejudice and polarization. Given their scale, using Large Language Models (LLMs) to assist expert counterspeech (CS) writing has gained interest, yet prior work has addressed these phenomena separately. We bridge this gap by studying CS generation in contexts where both hate and misinformation co-occur. We test three knowledge-driven generation strategies: first we prompt an LLM with fact-checkers' guidelines and fact-checking articles; secondly, with NGOs' guidelines and reports; thirdly, we create a mixed strategy that combines guidelines and documents from both. 23 experts revise the generated CS, which are assessed via human and automatic metrics. While LLMs produce adequate CS in 40% of cases, expert edits substantially improve naturalness, exhaustiveness, and adherence to guidelines. Based on the post-edited CS, the mixed strategy proves to be the most effective in crowdsourcing evaluation, pairing strong factual correction with stereotype mitigation and empathetic engagement. We release a dataset of hateful and misinformed claims with expert-verified CS and supporting knowledge.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper studies LLM-assisted counterspeech generation for online claims that combine hate speech and misinformation. It evaluates three knowledge-driven prompting strategies—fact-checker guidelines plus articles, NGO guidelines plus reports, and a mixed strategy combining both—on a new dataset of such claims. Generated outputs are revised by 23 experts and assessed with expert judgments, crowdsourced ratings, and automatic metrics. The authors report that LLMs produce adequate counterspeech in 40% of cases, that expert edits substantially improve naturalness, exhaustiveness, and guideline adherence, and that the mixed strategy yields the strongest post-edit performance in crowdsourced evaluation, combining factual correction with stereotype mitigation and empathy. A dataset of claims, expert-verified counterspeech, and supporting knowledge is released.

Significance. If the central claims hold after addressing evaluation confounds, the work fills a clear gap by treating hate speech and misinformation jointly rather than separately, and the released dataset constitutes a concrete, reusable resource for future research on assisted counterspeech. The empirical comparison of knowledge sources offers practical guidance for prompt design in high-stakes social-media settings. These contributions are proportionate to the modest scale of the study and would be strengthened by clearer isolation of generation effects from post-editing.

major comments (3)
  1. [Abstract] Abstract and evaluation sections: the claim that the mixed strategy is most effective rests on crowdsourced assessment of post-edited counterspeech. Because expert revisions are applied uniformly after generation and no pre-edit crowdsourced scores or per-strategy edit-distance statistics are reported, any advantage could arise from differential revision effort rather than from the prompting strategy itself. This attribution issue is load-bearing for the main result.
  2. [Evaluation] Evaluation and results sections: the abstract states a 40% adequate rate for LLM outputs and substantial expert improvements, yet no baselines, statistical significance tests, or complete metric definitions (e.g., exact criteria for “adequate,” “naturalness,” or “exhaustiveness”) are supplied. Without these, the quantitative claims cannot be properly interpreted or replicated.
  3. [Methods] Methods and data sections: the paper assumes that the chosen fact-checking articles and NGO reports supply representative, unbiased, and sufficient knowledge across diverse hate-plus-misinformation cases. No analysis of source selection criteria, potential coverage gaps, or bias checks is provided, which directly affects the validity of the knowledge-driven strategies.
minor comments (2)
  1. [Methods] Clarify the exact composition of the mixed strategy (e.g., how guidelines and documents from both sources are combined in the prompt) and report the number of claims per strategy to allow readers to assess balance.
  2. [Evaluation] Add a table or figure summarizing inter-annotator agreement for both expert and crowdsourced ratings.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. The comments highlight important issues around evaluation rigor and attribution that we will address in the revision. Below we respond point by point to the major comments.

read point-by-point responses

  1. Referee: [Abstract] Abstract and evaluation sections: the claim that the mixed strategy is most effective rests on crowdsourced assessment of post-edited counterspeech. Because expert revisions are applied uniformly after generation and no pre-edit crowdsourced scores or per-strategy edit-distance statistics are reported, any advantage could arise from differential revision effort rather than from the prompting strategy itself. This attribution issue is load-bearing for the main result.

    Authors: We agree this is a substantive concern that affects interpretation of the main result. In the revised manuscript we will add crowdsourced ratings of the pre-edited LLM outputs for each strategy and report per-strategy edit-distance statistics (or equivalent measures of revision effort) so that readers can assess whether performance differences are attributable to the prompting strategies rather than to uneven post-editing. revision: yes

  2. Referee: [Evaluation] Evaluation and results sections: the abstract states a 40% adequate rate for LLM outputs and substantial expert improvements, yet no baselines, statistical significance tests, or complete metric definitions (e.g., exact criteria for “adequate,” “naturalness,” or “exhaustiveness”) are supplied. Without these, the quantitative claims cannot be properly interpreted or replicated.

    Authors: We accept that the current presentation lacks necessary context for interpretation and replication. We will introduce a simple baseline (zero-shot generation without knowledge sources), add statistical significance testing for the key comparisons, and provide explicit annotation guidelines with criteria and examples for the labels “adequate,” “naturalness,” and “exhaustiveness.” revision: yes

  3. Referee: [Methods] Methods and data sections: the paper assumes that the chosen fact-checking articles and NGO reports supply representative, unbiased, and sufficient knowledge across diverse hate-plus-misinformation cases. No analysis of source selection criteria, potential coverage gaps, or bias checks is provided, which directly affects the validity of the knowledge-driven strategies.

    Authors: We recognize that explicit justification of the knowledge sources is required. The revision will include a new subsection describing the selection criteria for the fact-checking articles and NGO reports, acknowledging potential coverage gaps across topics and regions, and noting any known organizational perspectives. Full quantitative bias audits are beyond the scope of the current dataset but we will add a limitations discussion on this point. revision: partial

Circularity Check

0 steps flagged

Empirical evaluation of generation strategies is self-contained

full rationale

The paper presents an empirical workflow: curating hateful/misinformed claims, prompting an LLM under three distinct knowledge conditions (fact-checker guidelines+articles, NGO guidelines+reports, and their combination), obtaining expert post-edits, and then measuring the resulting counterspeech with fresh crowdsourced and automatic metrics. The central finding that the mixed strategy performs best is derived directly from these new human ratings rather than from any parameter fitted to the target outcome, any self-referential definition, or a load-bearing self-citation chain. No equations, uniqueness theorems, or ansatzes are invoked; the derivation therefore does not reduce to its inputs by construction and remains externally falsifiable through the released dataset and independent replication.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The work rests on standard NLP assumptions about prompt effectiveness and the validity of expert and crowdsourced judgments; no new entities or fitted parameters are introduced beyond the experimental setup.

axioms (1)
  • domain assumption Expert human revisions reliably improve naturalness, exhaustiveness, and guideline adherence of LLM-generated text.
    Invoked when comparing pre- and post-edited counterspeech quality.

pith-pipeline@v0.9.0 · 5722 in / 1190 out tokens · 42915 ms · 2026-05-22T07:02:39.739471+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

300 extracted references · 300 canonical work pages · 19 internal anchors

  1. [1]

    2024 , month =

    Charlotte Green , title =. 2024 , month =

    work page 2024

  2. [2]

    Proceedings of the ACM on Web Conference 2025 , pages=

    Contextualized counterspeech: Strategies for adaptation, personalization, and evaluation , author=. Proceedings of the ACM on Web Conference 2025 , pages=

    work page 2025

  3. [3]

    Transactions of the Association for Computational Linguistics , volume=

    Justilm: Few-shot justification generation for explainable fact-checking of real-world claims , author=. Transactions of the Association for Computational Linguistics , volume=. 2024 , publisher=

    work page 2024

  4. [4]

    Proceedings of the 18th International Natural Language Generation Conference , pages=

    Face the Facts! Evaluating RAG-based Pipelines for Professional Fact-Checking , author=. Proceedings of the 18th International Natural Language Generation Conference , pages=

    work page

  5. [5]

    (No Title) , year=

    The technique of clear writing , author=. (No Title) , year=

    work page

  6. [6]

    new media & society , volume=

    Global misinformation trends: Commonalities and differences in topics, sources of falsehoods, and deception strategies across eight countries , author=. new media & society , volume=. 2025 , publisher=

    work page 2025

  7. [7]

    2016 , month =

    Full Fact, Team , title =. 2016 , month =

    work page 2016

  8. [8]

    2023 , month =

    Josh Kelety , title =. 2023 , month =

    work page 2023

  9. [9]

    Realities: Bridging the gap between myths and realities about LGBTI people , howpublished =

    Myths vs. Realities: Bridging the gap between myths and realities about LGBTI people , howpublished =

    work page

  10. [10]

    , author=

    The proof and measurement of association between two things. , author=. 1961 , publisher=

    work page 1961

  11. [11]

    Biochemia medica , volume=

    Interrater reliability: the kappa statistic , author=. Biochemia medica , volume=. 2012 , publisher=

    work page 2012

  12. [12]

    , author=

    A new readability yardstick. , author=. Journal of applied psychology , volume=. 1948 , publisher=

    work page 1948

  13. [13]

    Derivation of new readability formulas (automated readability index, fog count and flesch reading ease formula) for navy enlisted personnel , author=

    work page

  14. [14]

    Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=

    End-to-end multimodal fact-checking and explanation generation: A challenging dataset and models , author=. Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages=

    work page

  15. [15]

    Accountable and Explainable Methods for Complex Reasoning over Text , pages=

    Generating fact checking explanations , author=. Accountable and Explainable Methods for Complex Reasoning over Text , pages=. 2024 , publisher=

    work page 2024

  16. [16]

    Understanding Counterspeech for Online Harm Mitigation

    Chung, Yi-Ling and Abercrombie, Gavin and Enock, Florence and Bright, Jonathan and Rieser, Verena. Understanding Counterspeech for Online Harm Mitigation. Northern European Journal of Language Technology. 2024. doi:10.3384/nejlt.2000-1533.2024.5203

    work page doi:10.3384/nejlt.2000-1533.2024.5203 2024

  17. [17]

    NLP for Counterspeech against Hate and Misinformation ( CSHAM )

    Russo, Daniel and Bonaldi, Helena and Chung, Yi-Ling and Abercrombie, Gavin and Guerini, Marco. NLP for Counterspeech against Hate and Misinformation ( CSHAM ). Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 5: Tutorial Abstracts). 2025. doi:10.18653/v1/2025.acl-tutorials.6

    work page doi:10.18653/v1/2025.acl-tutorials.6 2025

  18. [18]

    Proceedings of the twelfth ACM international conference on web search and data mining , pages=

    Exfakt: A framework for explaining facts over knowledge graphs and text , author=. Proceedings of the twelfth ACM international conference on web search and data mining , pages=

    work page

  19. [19]

    on the impact of pre-aggregation on the evaluation of highly subjective tasks , author=

    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 , author=. CEUR workshop proceedings , volume=. 2020 , organization=

    work page 2020

  20. [20]

    2022 , publisher=

    Digital work in the planetary market , author=. 2022 , publisher=

    work page 2022

  21. [21]

    2020 , publisher=

    Behind the screen: Content moderation in the shadows of social media , author=. 2020 , publisher=

    work page 2020

  22. [22]

    The Risk of Racial Bias in Hate Speech Detection

    Sap, Maarten and Card, Dallas and Gabriel, Saadia and Choi, Yejin and Smith, Noah A. The Risk of Racial Bias in Hate Speech Detection. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. 2019. doi:10.18653/v1/P19-1163

    work page doi:10.18653/v1/p19-1163 2019

  23. [23]

    Proceedings of the third workshop on abusive language online , year=

    Challenges and frontiers in abusive content detection , author=. Proceedings of the third workshop on abusive language online , year=

    work page

  24. [24]

    A systematic review of hate speech automatic detection using natural language processing , journal =

    Md Saroar Jahan and Mourad Oussalah , keywords =. A systematic review of hate speech automatic detection using natural language processing , journal =. 2023 , issn =. doi:https://doi.org/10.1016/j.neucom.2023.126232 , url =

    work page doi:10.1016/j.neucom.2023.126232 2023

  25. [25]

    Is Attention Interpretable?

    Serrano, Sofia and Smith, Noah A. Is Attention Interpretable?. Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics. 2019. doi:10.18653/v1/P19-1282

    work page doi:10.18653/v1/p19-1282 2019

  26. [26]

    A ttention is not E xplanation

    Jain, Sarthak and Wallace, Byron C. A ttention is not E xplanation. Proceedings of the 2019 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers). 2019. doi:10.18653/v1/N19-1357

    work page doi:10.18653/v1/n19-1357 2019

  27. [27]

    RU 22 F act: Optimizing Evidence for Multilingual Explainable Fact-Checking on R ussia- U kraine Conflict

    Zeng, Yirong and Ding, Xiao and Zhao, Yi and Li, Xiangyu and Zhang, Jie and Yao, Chao and Liu, Ting and Qin, Bing. RU 22 F act: Optimizing Evidence for Multilingual Explainable Fact-Checking on R ussia- U kraine Conflict. Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). 2024

    work page 2024

  28. [28]

    Proceedings of the 2020 Truth and Trust Online (TTO 2020) , pages=

    e-fever: Explanations and summaries for automated fact checking , author=. Proceedings of the 2020 Truth and Trust Online (TTO 2020) , pages=. 2020 , publisher=

    work page 2020

  29. [29]

    Findings of the Association for Computational Linguistics: ACL 2025 , pages=

    EuroVerdict: A Multilingual Dataset for Verdict Generation Against Misinformation , author=. Findings of the Association for Computational Linguistics: ACL 2025 , pages=

    work page 2025

  30. [30]

    D e C lar E : Debunking Fake News and False Claims using Evidence-Aware Deep Learning

    Popat, Kashyap and Mukherjee, Subhabrata and Yates, Andrew and Weikum, Gerhard. D e C lar E : Debunking Fake News and False Claims using Evidence-Aware Deep Learning. Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing. 2018. doi:10.18653/v1/D18-1003

    work page doi:10.18653/v1/d18-1003 2018

  31. [31]

    Proceedings of the 1st Workshop on CounterSpeech for Online Abuse (CS4OA) , pages=

    From generic to personalized: Investigating strategies for generating targeted counter narratives against hate speech , author=. Proceedings of the 1st Workshop on CounterSpeech for Online Abuse (CS4OA) , pages=

    work page

  32. [32]

    GPT-4 Technical Report

    Gpt-4 technical report , author=. arXiv preprint arXiv:2303.08774 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  33. [33]

    Pointer Sentinel Mixture Models

    Pointer sentinel mixture models , author=. arXiv preprint arXiv:1609.07843 , year=

    work page internal anchor Pith review Pith/arXiv arXiv

  34. [34]

    N ewsroom: A Dataset of 1.3 Million Summaries with Diverse Extractive Strategies

    Grusky, Max and Naaman, Mor and Artzi, Yoav. N ewsroom: A Dataset of 1.3 Million Summaries with Diverse Extractive Strategies. Proceedings of the 2018 Conference of the North A merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers). 2018. doi:10.18653/v1/N18-1065

    work page doi:10.18653/v1/n18-1065 2018

  35. [35]

    Towards Knowledge-Grounded Counter Narrative Generation for Hate Speech

    Chung, Yi-Ling and Tekiro g lu, Serra Sinem and Guerini, Marco. Towards Knowledge-Grounded Counter Narrative Generation for Hate Speech. Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021. 2021. doi:10.18653/v1/2021.findings-acl.79

    work page doi:10.18653/v1/2021.findings-acl.79 2021

  36. [36]

    Procesamiento del lenguaje natural , volume=

    Automatic counter-narrative generation for hate speech in spanish , author=. Procesamiento del lenguaje natural , volume=

    work page

  37. [37]

    Generating Counter Narratives against Online Hate Speech: Data and Strategies

    Tekiro g lu, Serra Sinem and Chung, Yi-Ling and Guerini, Marco. Generating Counter Narratives against Online Hate Speech: Data and Strategies. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020. doi:10.18653/v1/2020.acl-main.110

    work page doi:10.18653/v1/2020.acl-main.110 2020

  38. [38]

    Proceedings of the First Workshop on Multilingual Counterspeech Generation , pages=

    TrenTeam at Multilingual Counterspeech Generation: Multilingual Passage Re-Ranking Approaches for Knowledge-Driven Counterspeech Generation Against Hate , author=. Proceedings of the First Workshop on Multilingual Counterspeech Generation , pages=

    work page

  39. [39]

    A Benchmark Dataset for Learning to Intervene in Online Hate Speech

    Qian, Jing and Bethke, Anna and Liu, Yinyin and Belding, Elizabeth and Wang, William Yang. A Benchmark Dataset for Learning to Intervene in Online Hate Speech. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019. doi:10.18653/v...

    work page doi:10.18653/v1/d19-1482 2019

  40. [40]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    Pinpointing fine-grained relationships between hateful tweets and replies , author=. Proceedings of the AAAI Conference on Artificial Intelligence , volume=

    work page

  41. [41]

    Countering Misinformation via Emotional Response Generation

    Russo, Daniel and Kaszefski-Yaschuk, Shane and Staiano, Jacopo and Guerini, Marco. Countering Misinformation via Emotional Response Generation. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023. doi:10.18653/v1/2023.emnlp-main.703

    work page doi:10.18653/v1/2023.emnlp-main.703 2023

  42. [42]

    This is fake news

    “This is fake news”: Investigating the role of conformity to other users’ views when commenting on and spreading disinformation in social media , author=. Computers in Human Behavior , volume=. 2019 , publisher=

    work page 2019

  43. [43]

    Political Behavior , volume=

    When corrections fail: The persistence of political misperceptions , author=. Political Behavior , volume=. 2010 , publisher=

    work page 2010

  44. [44]

    2018 , school=

    (Dis) continuing the continued influence effect of misinformation , author=. 2018 , school=

    work page 2018

  45. [45]

    , author=

    Fear, misinformation, and innumerates: how the Wakefield paper, the press, and advocacy groups damaged the public health. , author=. Vaccine , volume=

    work page

  46. [46]

    The lancet , volume=

    RETRACTED: Ileal-lymphoid-nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children , author=. The lancet , volume=. 1998 , publisher=

    work page 1998

  47. [47]

    Memory & cognition , volume=

    Explicit warnings reduce but do not eliminate the continued influence of misinformation , author=. Memory & cognition , volume=. 2010 , publisher=

    work page 2010

  48. [48]

    Studies in Communication and Media , number=

    Countering misinformation: Strategies, challenges, and uncertainties , author=. Studies in Communication and Media , number=. 2019 , publisher=

    work page 2019

  49. [49]

    Proceedings of the 15th ACM Web Science Conference 2023 , pages=

    Characterizing and predicting social correction on twitter , author=. Proceedings of the 15th ACM Web Science Conference 2023 , pages=

    work page 2023

  50. [50]

    2014 , publisher=

    How people read on the web: The eyetracking evidence , author=. 2014 , publisher=

    work page 2014

  51. [51]

    Liar, Liar Pants on Fire

    Wang, William Yang. ``Liar, Liar Pants on Fire'': A New Benchmark Dataset for Fake News Detection. Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 2017. doi:10.18653/v1/P17-2067

    work page doi:10.18653/v1/p17-2067 2017

  52. [52]

    Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    Language-aware truth assessment of fact candidates , author=. Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages=

    work page

  53. [53]

    Management science , volume=

    The implied truth effect: Attaching warnings to a subset of fake news headlines increases perceived accuracy of headlines without warnings , author=. Management science , volume=. 2020 , publisher=

    work page 2020

  54. [54]

    Proceedings of the International AAAI Conference on Web and Social Media , volume=

    Do explanations increase the effectiveness of AI-crowd generated fake news warnings? , author=. Proceedings of the International AAAI Conference on Web and Social Media , volume=

    work page

  55. [55]

    30th USENIX Security Symposium (USENIX Security 21) , pages=

    Adapting security warnings to counter online disinformation , author=. 30th USENIX Security Symposium (USENIX Security 21) , pages=

    work page

  56. [56]

    Computers & security , volume=

    Misinformation warnings: Twitter’s soft moderation effects on COVID-19 vaccine belief echoes , author=. Computers & security , volume=. 2022 , publisher=

    work page 2022

  57. [57]

    Transactions of the Association for Computational Linguistics , volume=

    A survey on automated fact-checking , author=. Transactions of the Association for Computational Linguistics , volume=. 2022 , publisher=

    work page 2022

  58. [58]

    Communication Research Reports , volume=

    Playing nice: Modeling civility in online political discussions , author=. Communication Research Reports , volume=. 2015 , publisher=

    work page 2015

  59. [59]

    Personality and Social Psychology Bulletin , volume=

    Further progress in understanding the effects of derogatory ethnic labels: The role of preexisting attitudes toward the targeted group , author=. Personality and Social Psychology Bulletin , volume=. 1996 , publisher=

    work page 1996

  60. [60]

    Proceedings of the international AAAI conference on web and social media , volume=

    If you have a reliable source, say something: effects of correction comments on covid-19 misinformation , author=. Proceedings of the international AAAI conference on web and social media , volume=

    work page

  61. [61]

    Frontiers in Psychology , volume=

    Changing conspiracy beliefs through rationality and ridiculing , author=. Frontiers in Psychology , volume=. 2016 , publisher=

    work page 2016

  62. [62]

    American politics research , volume=

    Debating the truth: The impact of fact-checking during electoral debates , author=. American politics research , volume=. 2017 , publisher=

    work page 2017

  63. [63]

    2015 , publisher=

    Hate speech law: A philosophical examination , author=. 2015 , publisher=

    work page 2015

  64. [64]

    Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems , pages=

    Counterspeakers’ perspectives: Unveiling barriers and ai needs in the fight against online hate , author=. Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems , pages=

    work page 2024

  65. [65]

    IEEE INFOCOM 2021-IEEE conference on computer communications , pages=

    Setting the record straighter on shadow banning , author=. IEEE INFOCOM 2021-IEEE conference on computer communications , pages=. 2021 , organization=

    work page 2021

  66. [66]

    Proceedings of the ACM on human-computer interaction , volume=

    You can't stay here: The efficacy of reddit's 2015 ban examined through hate speech , author=. Proceedings of the ACM on human-computer interaction , volume=. 2017 , publisher=

    work page 2015

  67. [67]

    PNAS nexus , volume=

    Deplatforming did not decrease Parler users’ activity on fringe social media , author=. PNAS nexus , volume=. 2023 , publisher=

    work page 2023

  68. [68]

    Internet Policy Review , volume=

    Expanding the debate about content moderation: Scholarly research agendas for the coming policy debates , author=. Internet Policy Review , volume=. 2020 , publisher=

    work page 2020

  69. [69]

    Information, Communication & Society , volume=

    Civilized truths, hateful lies? Incivility and hate speech in false information--evidence from fact-checked statements in the US , author=. Information, Communication & Society , volume=. 2022 , publisher=

    work page 2022

  70. [70]

    arXiv preprint arXiv:2404.08110 , year=

    Toxic Synergy Between Hate Speech and Fake News Exposure , author=. arXiv preprint arXiv:2404.08110 , year=

    work page arXiv

  71. [71]

    Psychological science in the public interest , volume=

    Misinformation and its correction: Continued influence and successful debiasing , author=. Psychological science in the public interest , volume=. 2012 , publisher=

    work page 2012

  72. [72]

    Naive realism

    Actual versus assumed differences in construal:" Naive realism" in intergroup perception and conflict. , author=. Journal of personality and social psychology , volume=. 1995 , publisher=

    work page 1995

  73. [73]

    The Routledge companion to media disinformation and populism , pages=

    Global responses to misinformation and populism , author=. The Routledge companion to media disinformation and populism , pages=. 2021 , publisher=

    work page 2021

  74. [74]

    NIM Marketing Intelligence Review , volume=

    How truthiness, fake news and post-fact endanger brands and what to do about it , author=. NIM Marketing Intelligence Review , volume=. 2018 , publisher=

    work page 2018

  75. [75]

    International Journal of Environmental Research and Public Health , volume=

    The determinants of panic buying during COVID-19 , author=. International Journal of Environmental Research and Public Health , volume=. 2021 , publisher=

    work page 2021

  76. [76]

    post-truth

    Beyond misinformation: Understanding and coping with the “post-truth” era , author=. Journal of applied research in memory and cognition , volume=. 2017 , publisher=

    work page 2017

  77. [77]

    Communication Research , volume=

    Credibility perceptions and detection accuracy of fake news headlines on social media: Effects of truth-bias and endorsement cues , author=. Communication Research , volume=. 2022 , publisher=

    work page 2022

  78. [78]

    Perspectives on Psychological Science , volume=

    (Why) is misinformation a problem? , author=. Perspectives on Psychological Science , volume=. 2023 , publisher=

    work page 2023

  79. [79]

    science , volume=

    The spread of true and false news online , author=. science , volume=. 2018 , publisher=

    work page 2018

  80. [80]

    Personality and Social Psychology Review , volume=

    The truth about the truth: A meta-analytic review of the truth effect , author=. Personality and Social Psychology Review , volume=. 2010 , publisher=

    work page 2010

Showing first 80 references.