Can the current generation of large language models (LLMs) produce an adequate book index?

Artificial intelligence (AI) is seen as a threat to many creative professions and book indexers are no exception. This paper supplements our previously published article by investigating the capacity of the current generation of large language models (LLMs) to index a book chapter. The previous article found that an earlier generation of LLMs performed poorly on the three criteria of completeness, navigability, and accuracy. While the current generation does not under-index as severely, it has issues with excessive subheadings and navigability, and continues to fail a standard indexing test for accuracy. LLMs still cannot produce an adequate index.