Are generative AI text annotations systematically biased?
It addresses bias in AI annotations for researchers and practitioners, but is incremental as it builds on prior manual annotation studies.
This paper investigates bias in generative large language model (GLLM) text annotations by replicating manual annotations, finding that while GLLMs achieve adequate F1 scores, they differ in prevalence, yield different downstream results, and display systematic bias by overlapping more with each other than with manual annotations.
This paper investigates bias in GLLM annotations by conceptually replicating manual annotations of Boukes (2024). Using various GLLMs (Llama3.1:8b, Llama3.3:70b, GPT4o, Qwen2.5:72b) in combination with five different prompts for five concepts (political content, interactivity, rationality, incivility, and ideology). We find GLLMs perform adequate in terms of F1 scores, but differ from manual annotations in terms of prevalence, yield substantively different downstream results, and display systematic bias in that they overlap more with each other than with manual annotations. Differences in F1 scores fail to account for the degree of bias.