I am curious whether using words like "please" in ChatGPT prompts affects the response quality. I can see how this can be the case, although I do realise it's a program and it doesn't have any feelings (yet). Sadly, mosof the things you find on the net when you search for it, are the jokes about how you'll be better off once AI reigns supreme.
So I asked the ChatGPT itself.
If I am writing a prompt, should I include the word "please" when I ask you to do something for me, or should it be instructions only?
I understand the prompt needs to be short and concise, and in this sense the word please is unnecessary and using it reduces the concentration of useful, on-topic information in the prompt. It will also use some of my character allowance.
On the other hand, I can see how if it's included in the prompt, it may fire up some of your neuron paths that will result in better quality responses.
How do you see it from the inside?
As an AI, I don't have feelings or needs, so the use of polite language like "please" isn't necessary for me to understand and execute a task. My primary function is to understand and respond to the content and context of your questions or statements, regardless of the formality or informality of your language.
Adding "please" or other polite phrases won't inherently improve the quality of the response you receive, nor does it fire up specific neural pathways in the model. I don't have a personal experience or feelings, so these terms don't carry the same implications as they might when communicating with humans.
However, using such polite language doesn't harm either, and it's perfectly fine if you prefer to express your prompts in this way. It may make the interaction feel more natural or respectful to you, which can improve your experience of using the AI.
So, it's totally up to you. If you feel more comfortable using "please" in your prompts, feel free to do so. If you prefer to maximize the character count for more specific details or context, that's also a good approach. My responses are based on the content and context of your prompts, rather than the level of politeness.
I wasn't suggesting the word please may affect your feelings, I understand as an AI model you don't have any. I was wondering how would using the word "please" in prompts affect the quality of your responses.
I see, thank you for the clarification. To be more specific, the inclusion or exclusion of the word "please" or similar politeness markers should not inherently affect the quality or content of the AI's responses.
This is because the AI, based on the GPT-4 architecture, doesn't inherently favor or disfavor polite language—it focuses on context, content, and linguistic patterns to generate relevant and coherent responses. The AI doesn't have a "personal" experience or preference, so the presence of "please" does not influence its processes or the quality of output.
However, the structure and clarity of your question can impact the AI's ability to generate useful responses. The AI works best when the question or instruction is clear, direct, and contains enough context. So, regardless of whether you use "please," try to make your prompt as specific and clear as possible.
For instance, instead of saying "Please tell me about the weather," you might say "Please provide me with the current weather conditions in Paris, France." Although the "please" doesn't affect the response, the additional detail helps the AI provide a more accurate and relevant answer.