ChatGPT vs Bard: Which is a Better Writer?

Abstract

The emergence and subsequent widespread adoption of Artificial Intelligence (AI) Language models such as ChatGPT and Google Bard has sparked intense interest in the capabilities and applications of AI language models in education and language learning. However, not much is known about how AI language models perform linguistically. This study examined the effects of AI language models, ChatGPT and Bard, on the writing quality and linguistic features (i.e., lexical sophistication, syntactic complexity, and cohesion) of AI written output. Data was collected from both AI language models, which were tasked to answer a total of 120 essay prompts on topics ranging from Arts and Humanities, Social Sciences, and Science. The responses generated from both AI language models were analyzed for their writing quality, lexical sophistication, syntactic complexity, and cohesion. Results revealed that ChatGPT is stronger in producing higher quality text which uses more sophisticated words and complex sentences, whereas Bard is stronger in creating more cohesive texts. Results of the study are discussed in the context of AI language models’ training methods, and implications for using AI in education, as well as, language teaching and learning.

Ähnliche Arbeiten