CoChat

Generative AI Search Engines as Arbiters of Public Knowledge: An Audit of Bias and Authority

Li, Alice, Sinnamon, Luanne · 2024 · Proceedings of the Association for Information Science and Technology
Cited by 15

The DOI matches its Crossref record; 17 of the 17 references it cites were found in a registry or an index.

Abstract

This paper reports on an audit study of generative AI systems (ChatGPT, Bing Chat, and Perplexity), which investigates how these new search engines construct responses and establish authority for topics of public importance. We collected system responses using a set of 48 authentic queries for 4 topics over 7 days, and analyzed the data using sentiment analysis, inductive coding, and source classification. Results provide an overview of the nature of system responses across these systems, and provide evidence of sentiment bias based on the queries and topics, and commercial and geographic biases in sources. The quality of sources used to support claims is uneven, relying heavily on News and Media, Business and Digital Media websites. Implications for system users emphasize the need to critically examine Generative AI system outputs when making decisions related to public interest and personal well‐being.

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.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1002/pra2.1021

Cites

17 of 17 references verified0 citation errors · 0 not checkable

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.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

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

Research works printed but not linked: 40

  1. 10.17705/1jais.00458 · doi:10.17705/1jais.00458
  2. 10.1145/3406522.3446023 · DOI · Paper page
  3. Bubeck S. Chandrasekaran V. Eldan R. Gehrke J. Horvitz E. Kamar E. …Zhang Y.(2023).Sparks of artificial general intelligence: Early experiments with gpt‐4. arXiv preprint.https://doi.org/10.48550/arXiv.2303.12712 · DOI · Paper page
  4. Capra R. &Arguello J.(2023).How does AI chat change search behaviors?arXiv preprint.https://doi.org/10.48550/arXiv.2307.03826 · DOI
  5. 10.1145/1863879.1863883 · DOI
  6. 10.3389/fmed.2020.00400 · DOI
  7. 10.4324/9780429448546 · DOI · Paper page
  8. 10.1038/s41586-020-2649-2 · DOI · Paper page
  9. Hovland C., 1953, Communication and persuasion
  10. 10.1145/3392854 · DOI · Paper page
  11. 10.1145/3331184.3331314 · DOI
  12. 10.1002/pra2.927 · DOI
  13. 10.4135/9781071878781.n13 · doi:10.4135/9781071878781.n13
  14. 10.1007/978-3-031-09316-6_5 · DOI
  15. Li A. &Sinnamon L.(2023 August 28–September 2).Examining query sentiment bias effects on search results in large language models.The Symposium on Future Directions in Information Access (FDIA) co‐located with the 2023 European Summer School on Information Retrieval (ESSIR).https://2023.essir.eu/FDIA/papers/FDIA_2023_paper_2.pdf
  16. Li D. Sun Z. Hu X. Liu Z. Chen Z. Hu B. …Zhang M.(2023).A survey of large language Models attribution. arXiv preprint.https://doi.org/10.48550/arXiv.2311.03731 · DOI
  17. 10.1177/20539517231191528 · DOI
  18. McKinney W.(2010).Data structures for statistical computing in Python.Proceedings of the Python in Science Conferences.https://doi.org/10.25080/majora-92bf1922-00a · DOI · Paper page
  19. Mehdi Y., 2023, Confirmed: the new Bing runs on OpenAI's GPT‐4
  20. 10.1145/3449100 · DOI
  21. 10.1561/1100000083 · DOI · Paper page
  22. 10.18574/nyu/9781479833641.001.0001 · DOI · Paper page
  23. 10.1177/160940691989922 · doi:10.1177/160940691989922
  24. 10.1002/asi.10017 · DOI · Paper page
  25. 10.1145/3627508.3638345 · DOI
  26. 10.1108/JD-12-2018-0209 · DOI · Paper page
  27. Sandvig C., 2014, Auditing algorithms: Research methods for detecting discrimination on internet platforms, Data and discrimination: converting critical concerns into productive inquiry, 22, 4349
  28. 10.1145/3498366.3505816 · DOI · Paper page
  29. Shi X. Liu J. Liu Y. Cheng Q. &Lu W.(2023).Know where to go: Make LLM a relevant responsible and trustworthy searcher. arXiv preprint.https://doi.org/10.48550/arXiv.2310.12443 · DOI
  30. Smith M. S., 2024, Perplexity.ai revamps Google SEO model for LLM era > AI search leader mixes Meta‐built smarts with scrappy startup fervor, IEEE Spectrum
  31. 10.1002/asi.20652 · DOI · Paper page
  32. 10.1108/00220411111164709 · DOI
  33. Taraborelli D., 2008, Proceeding of the 2008 Conference on Current Issues in Computing and Philosophy, 194
  34. 10.1177/1098214005283748 · DOI · Paper page
  35. 10.1609/aaai.v37i11.26599 · DOI
  36. Urman A. &Makhortykh M.(2023).The silence of the LLMs: Cross‐lingual analysis of political bias and false information prevalence in ChatGPT Google Bard and Bing Chat. OSF Preprints.https://doi.org/10.31219/osf.io/q9v8f · DOI · Paper page
  37. 10.1145/3627508.3638305 · DOI
  38. 10.1145/2484028.2484053 · DOI · Paper page
  39. Wilson P., 1983, Second‐hand knowledge: An inquiry into cognitive authority
  40. 10.1145/3627508.3638316 · DOI

Cited by

See all 15

Metadata

Authors
Li, A., & Sinnamon, L.
Affiliations
The University of British Columbia Canada
Venue
Proceedings of the Association for Information Science and Technology
Publisher
Wiley
Published
2024-10
Volume
61
Issue
1
Pages
205-217
ISSN
2373-9231
Cited by
15
References
17
Language
en

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

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.