Bullshit, Pragmatic Deception, and Natural Language Processing

2026-04-19 · Source: Paper Index on ACL Anthology · Field: Technology & Digital — Artificial Intelligence & Machine Learning · Depth: Expert, quick

Summary

The natural language processing (NLP) community has focused on fact-checking and fake news detection, but other forms of misinformation, such as "bullshit," remain largely unaddressed. Bullshit, defined as deception about one's goals rather than factual content, differs fundamentally from lying, making traditional fact-checking ineffective. This paper introduces "bullshitology" to NLP by proposing a QUD-based definition, outlining two distinct approaches for annotating bullshit, and identifying specific NLP methods suitable for classifying various types of linguistic bullshit. The work emphasizes detecting breaches in pragmatic conventions with deceptive intent, rather than verifying factual claims.

Key takeaway

For research scientists developing misinformation detection systems, you should recognize that traditional fact-checking is insufficient for "bullshit" because it concerns deceptive intent about goals, not factual accuracy. Your efforts should shift towards identifying breaches in pragmatic linguistic conventions to effectively classify this distinct form of misinformation, potentially integrating QUD-based analysis into your models.

Key insights

Bullshit deceives about goals, not facts, requiring pragmatic analysis in NLP.

Principles

• Bullshitters deceive about goals, not content.
• Fact-checking is useless against bullshit.
• Deception involves breaking pragmatic conventions.

Method

The paper proposes a QUD-based definition for bullshit, outlines two annotation approaches, and suggests NLP methods for classifying linguistic bullshit types.

In practice

• Develop NLP models for pragmatic convention breaches.
• Annotate datasets based on QUD-defined bullshit.
• Distinguish bullshit from factual lies in detection.

Topics

• Natural Language Processing
• Pragmatic Deception
• Bullshitology
• Misinformation Detection
• QUD-based Definition

Best for: Research Scientist, AI Scientist, NLP Engineer

Related on AIssential

• BiCon-Gate: Consistency-Gated De-colloquialisation for Dialogue Fact-Checking — BiCon-Gate improves dialogue fact-checking by consistently de-colloquializing claims with contextual semantic gating.
• Misinformation Span Detection in Videos via Audio Transcripts — Pinpointing misinformation within videos via audio transcript spans offers more granular, interpretable detection than video-level classification.
• Socially Responsible and Explainable Automated Fact-Checking and Hate Speech Detection — Explainable NLP methods and high-quality datasets can jointly optimize performance and interpretability for Portuguese hate speech and fact-checking.
• Inside BullshitBench: AI Models and Nonsense Detection — AI models often confidently answer nonsensical questions, but specific training can significantly improve their ability to detect and push back against flawed premises.
• AI put "synthetic quotes" in his book. But this author wants to keep using it. — AI tools offer efficiency but introduce insidious factual inaccuracies, challenging traditional verification processes and tempting even wary users with their "magical" utility.
• The fake disease that fooled the internet — and what it says about all of us — Human biases and reliance on superficial cues often lead to misjudging credibility, even with AI amplification of falsehoods.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Paper Index on ACL Anthology.