Tools like OpenAI’s ChatGPT chatbot stand out as revolutionary technologies that will make employees’ jobs extremely easy, and could even take away people’s jobs in the future. Many workers also rely on artificial intelligence to make their jobs easier. On the other hand, according to a new study ChatGPTquestions about programming by 52% gives wrong answer.

This is evident from a study conducted by researchers at Purdue University. 517 programming questions sourced from Stack Overflow ChatGPT was asked. The percentage of the artificial intelligence chatbot that answered these questions correctly was then measured. In their article, the researchers write “Our analysis showed us that 52% of ChatGPT responses contained misinformation and 77% were verbose. Yet, our study participants preferred 35% to use the ChatGPT responses due to the completeness and fluency of the ChatGPT responses. responses.” he said.

Developers failed to catch incorrect answers

The most disturbing part is that of the developers don’t always get wrong answers happened. According to the study, participants ignored incorrect information in ChatGPT responses 39% of the time. This means taking precautions against ChatGPT’s incorrect answers, the researchers said, and reveals the risk of apparently correct answers.

Of course, this study is just a single scenario chatbots too It can be adapted for different purposes. Companies that produce artificial intelligence tools are also continuing to work to ensure bots can provide more accurate answers.