In the world of artificial intelligence, there is often talk about which model is smarter or more technologically advanced. Less attention is paid to how AI communicates with the user. Meanwhile, the way a conversation is conducted can have a huge impact on the perception of the entire system. A journalist from TechRadar decided to conduct a simple experiment to see if ChatGPT could imitate the operating style of Google Gemini. It took just one properly constructed prompt. The results showed that the personality of AI can be just as important as its actual capabilities.
One prompt changed the character of ChatGPT
The author of the experiment instructed ChatGPT to respond more like Gemini. The model was to be more analytical, organised, and restrained, while also being less emotional and less conversational. The effects were immediate. The responses that previously resembled a conversation with a helpful advisor began to sound significantly more formal and academic. ChatGPT started to employ balanced phrasing more frequently, analyse various scenarios, and avoid emotional language. In many cases, the responses were very similar to those generated by Gemini.
The style of conversation is of great importance
The experiment showed that users often evaluate AI models not only through the lens of knowledge or the quality of responses. Equally significant is the way information is presented. ChatGPT usually comes across as more human and engaged in the conversation. In contrast, Gemini tends to rely more on cool analysis and structured explanations. After changing the communication style, the author noticed that they began to perceive the same model differently, even though its capabilities remained unchanged. This suggests that user preferences may stem more from the nature of the conversation than from actual technological differences.
A simple experiment demonstrated that the personality of artificial intelligence has a significant impact on user experience. ChatGPT was able to effectively mimic the style of Gemini, which proves that we often do not compare the AI models themselves, but rather the way they communicate with us.
source: techradar
Redakcja Choose TV












