The American Society for Indexing held its 2025 Virtual Conference, “Spring into Action,” on Friday, April 25, 2025, and Saturday, April 26, 2025. Sessions were held virtually on Zoom each day.
In the first session on Saturday, Indexing and Generative AI, presenter Tanya Izzard gave an overview of how generative artificial intelligence (AI) tools work and shared examples of how they perform at indexing and other related tasks. She discussed some of the problems involved with using generative AI for indexing work, including reliability and legal and contractual issues.
She gave four examples of how to use an AI tool, such as Claude, to give a text summary, index terms suggestion, create a list of names, and draft an index. In the text summary, she found that AI gave a reasonably accurate summary, with dates and names recorded, but it was fairly long. With the index terms suggestion and list of names, AI picked up some minor mentions and put in glosses. The index that AI produced wasn’t useful, and she said that she would save any time creating it this way.
Problems and issues with using AI include legal and contractual, such as normal restrictions on sharing proofs and manuscripts, reliability and accuracy, involving AI’s hallucinations and incorrect predictions, human versus algorithmic judgment, environment and ethical considerations, such as the high use of water, energy, land, and natural resources, and cost.
She concluded by saying that AI tools may speed up some indexing-related tasks, but cannot yet create an index and their actual contribution to indexing productivity seems doubtful.
In the next blog posting, I will discuss the next Saturday session from the ASI 2025 Virtual Conference. For more information about the services provided by the author of this blog, see the Stellar Searches LLC website, https://www.stellarsearches.com.