Many people find it surprising that AI is not yet sophisticated enough to produce a good index for a book--or, indeed, for a paper of any length. The problems with AI are problems that affect other functions, listed below.

 

Statement from the American Society for Indexing:

Using large language model AIs (LLMs) such as ChatGPT to generate indexes does not produce results that come anywhere near meeting our standards for excellence in Indexing. We also advise that documents not be uploaded or shared with LLMs without explicit client permission.

 

These LLMs fail at the indexer’s primary task: to ensure readers find needed information. Tests have shown that LLMs typically under-index book-length works, do not provide adequate structure or cross-references, and can insert false information (hallucinations) into the index.* Each of these is problematic:

• Severe under-indexing, with as few as 20-40% of the access points of human-generated indexes, can prevent readers from locating desired information and mislead them into thinking that omitted information isn’t in the book at all

• Absent/near-absent index structure, especially cross-references, prevents the reader from effectively navigating to subtopics and related topics while misrepresenting the focus of the book

• Hallucinations (including invented page references and even wholly nonexistent topics) waste the reader’s time and break the trust between the reader and the book

 

Future developments in AI may bring improvements, but at present we conclude the human brain of a professional indexer is still the best tool for analyzing, writing, and editing an index in full awareness of context as per quality standards.

​

*Bartmess, Elizabeth, “AI: Where You Can Use It and Where You Shouldn’t,” May 31, 2025, ISC/SCI Conference 2025, “Location! Location! Location.”