CoChat

Human heuristics for AI-generated language are flawed

Open access
Jakesch, Maurice, Hancock, Jeffrey T., Naaman, Mor · 2023 · Proceedings of the National Academy of Sciences

The DOI matches its Crossref record; 54 of the 54 references it cites were found in a registry or an index.

Abstract

Human communication is increasingly intermixed with language generated by AI. Across chat, email, and social media, AI systems suggest words, complete sentences, or produce entire conversations. AI-generated language is often not identified as such but presented as language written by humans, raising concerns about novel forms of deception and manipulation. Here, we study how humans discern whether verbal self-presentations, one of the most personal and consequential forms of language, were generated by AI. In six experiments, participants (N = 4,600) were unable to detect self-presentations generated by state-of-the-art AI language models in professional, hospitality, and dating contexts. A computational analysis of language features shows that human judgments of AI-generated language are hindered by intuitive but flawed heuristics such as associating first-person pronouns, use of contractions, or family topics with human-written language. We experimentally demonstrate that these heuristics make human judgment of AI-generated language predictable and manipulable, allowing AI systems to produce text perceived as “more human than human.” We discuss solutions, such as AI accents, to reduce the deceptive potential of language generated by AI, limiting the subversion of human intuition.

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 · 10.1073/pnas.2208839120

Cites

54 of 54 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1. Participants could not detect self-presentations generated by the current AI language models beyond chance in the three main experiments. Error bars represent 95% CIs for 6,000 to 16,000 judgments of 2,000 to 3,000 self-presentations per bar. Across three social contexts, discernment remained close to chance. Providing monetary incentives for accurate answers or telling participants whether their answers were correct did not increase accuracy.
Fig. 1. Participants could not detect self-presentations generated by the current AI language models beyond chance in the three main experiments. Error bars represent 95% CIs for 6,000 to 16,000 judgments of 2,000 to 3,000 self-presentations per bar. Across three social contexts, discernment remained close to chance. Providing monetary incentives for accurate answers or telling participants whether their answers were correct did not increase accuracy.CC BY-NC-ND · PubMed Central
Fig. 2. The three validation experiments show that AI systems can exploit humans’ flawed heuristics to generate optimized self-presentations (rightmost in each pane) more likely to be perceived as human than human-written self-presentations (leftmost) and regular AI-generated profiles (center). Error bars represent 95% CIs for 350 to 450 judgments of 100 self-presentations per bar.
Fig. 2. The three validation experiments show that AI systems can exploit humans’ flawed heuristics to generate optimized self-presentations (rightmost in each pane) more likely to be perceived as human than human-written self-presentations (leftmost) and regular AI-generated profiles (center). Error bars represent 95% CIs for 350 to 450 judgments of 100 self-presentations per bar.CC BY-NC-ND · PubMed Central

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?

6 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

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

Show all 54 references

Research works printed but not linked: 3

  1. A. Radford , Language models are unsupervised multitask learners. OpenAI Blog 1, 9 (2019).
  2. A. Nelson, S. Friedler, F. Fields-Meyer, Blueprint for an AI bill of rights: A vision for protecting our civil rights in the algorithmic age. White House Office of Science and Technology Policy (2022). (18 October 2022).
  3. J. Williams, Should AI always identify itself? It’s more complicated than you might think. Electronic Frontier Foundation (2018). (18 October 2022).

Cited by

See all 609

Metadata

Authors
Jakesch, M., Hancock, J. T., & Naaman, M.
Affiliations
  • Cornell University (US)
  • Stanford University (US)
Corresponding author
Jakesch, Maurice
Venue
Proceedings of the National Academy of Sciences
Published
2023-03-07
Volume
120
Issue
11
Pages
e2208839120-e2208839120
ISSN
  • 0027-8424 (ISSN-L)
  • 1091-6490
OpenAlex
W4323347737
Cited by
342
References
46
Citations by year
2026: 138 · 2025: 102 · 2024: 69 · 2023: 32
FWCI
32.10
Type
journal article
Language
en
Open access
Locations
  • Proceedings of the National Academy of Sciences (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
National Science Foundation (1901151/1901329)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Authorship Attribution and Profiling
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
  • Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
Keywords
AI-generated text detection, heuristics, intuitive judgment, AI language models, manipulation, first-person pronouns, contractions, human-AI interaction
MeSH terms
Heuristics, Artificial Intelligence, Choline O-Acetyltransferase, Communication, Humans, Language

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: Human heuristics for AI-generated language are flawed. https://app.cochat.ai/papers/10.1073/pnas.2208839120
The paper
Jakesch, M., Hancock, J. T., & Naaman, M. (2023). Human heuristics for AI-generated language are flawed. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.2208839120

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.