ChatGPT Answers More Than Half of Software Engineering Questions Incorrectly
ZDNET
11
Translated Data:
A study conducted by Purdue University has found that ChatGPT's accuracy in answering software engineering questions is low, with more than half of the responses being incorrect. Despite the comprehensive nature of the responses, users find it difficult to correctly identify the misinformation due to the intricate language used in the answers. The research indicates that ChatGPT's reliability in answering software engineering questions is inferior to traditional resources like Stack Overflow. Therefore, when seeking help in software engineering, it is advisable to still prefer traditional resources such as Stack Overflow.
© Copyright AIbase Base 2024, Click to View Source - https://www.aibase.com/news/279