Artificial intelligence appears to favor its own creative ideas over human ones, but it’s actually just distracted by how many fancy words it uses.
April 25, 2026
Original Paper
The Effect of Idea Elaboration on the Automatic Assessment of Idea Originality
arXiv · 2604.20569
The Takeaway
Self-preference bias in large language models is a result of verbosity rather than a love for AI-generated content. When AI judges the originality of an idea, it consistently gives higher scores to the more detailed and elaborated descriptions typical of machine output. Once the level of detail is equalized between human and AI ideas, this perceived bias completely disappears. This proves that we can fix AI ego simply by changing how ideas are presented. It also warns us that human judges might be making the same mistake by confusing a long explanation with a truly original thought.
From the abstract
Automatic systems are increasingly used to assess the originality of responses in creative tasks. They offer a potential solution to key limitations of human assessment (cost, fatigue, and subjectivity), but there is preliminary evidence of a self-preference bias. Accordingly, automatic systems tend to prefer outcomes that are more closely related to their style, rather than to the human one. In this paper, we investigated how Large Language Models (LLMs) align with human raters in assessing the