ChatGPT to Replace Crowdsourcing of Paraphrases for Intent Classification: Higher Diversity and Comparable Model Robustness
This work addresses the problem of reducing reliance on human labor for data generation in NLP, showing potential cost savings and efficiency gains, though it is incremental as it builds on existing crowdsourcing methods.
The study investigated whether ChatGPT can replace human crowdsourcing for generating paraphrases in intent classification, finding that ChatGPT-produced paraphrases are more diverse and yield models with comparable or better robustness.
The emergence of generative large language models (LLMs) raises the question: what will be its impact on crowdsourcing? Traditionally, crowdsourcing has been used for acquiring solutions to a wide variety of human-intelligence tasks, including ones involving text generation, modification or evaluation. For some of these tasks, models like ChatGPT can potentially substitute human workers. In this study, we investigate whether this is the case for the task of paraphrase generation for intent classification. We apply data collection methodology of an existing crowdsourcing study (similar scale, prompts and seed data) using ChatGPT and Falcon-40B. We show that ChatGPT-created paraphrases are more diverse and lead to at least as robust models.