Study evaluates NLP approaches for extracting keywords from crowdsourced collections

A project using the University of Oxford's Their Finest Hour Online Archive evaluated three Natural Language Processing approaches—Named Entity Recognition, Keyword Extraction, and Topic Modelling—to automate keyword extraction at scale. The study tested these methods across a range of techniques, from traditional statistical models to modern generative AI neural networks.

• Quantitative and qualitative findings indicate that NLP offers real potential for scaling keyword extraction in crowdsourced collections.
• No single method provides a complete solution, and model choice significantly shapes the results.
• Open-weight, extractive models are identified as best suited for responsible deployment.
• Generative AI introduces accountability risks due to its abstractive nature.

The authors argue that automated keyword extraction in crowdsourced collections raises distinct stewardship responsibilities regarding metadata produced by living contributors.