Key answer
Pronouns specify who does what; politeness expresses tone. Neither is a required prompt component. “Summarize the article below” and “Please summarize the article below” both work as requests. Clarify “I,” “you” and “we” when responsibilities, authorship or writing on someone's behalf matter. See A011 for more on pronouns.
Full explanation
Politeness can affect output, but research provides no universal rule that more courteous or harsher wording is better. Yin et al. tested eight politeness levels in English, Chinese and Japanese using GPT-3.5, GPT-4 and language-specific models. Extremely impolite versions often performed worse, while excessive politeness was not guaranteed to be best. Optimal levels varied by language and model; summarization changes could concern length as well as quality. Paper
Conversely, the short 2025 preprint *Mind Your Tone* reported 84.8% average accuracy for “very rude” GPT-4o prompts and 80.8% for “very polite” ones on 50 English multiple-choice questions. Questions and reference answers were AI-generated, and the main results concerned one model. This cannot establish that insulting AI makes it smarter. Text and Table 2
Beginners can use natural, direct, ordinarily polite language. “Please summarize” is a convenient default, not a scientifically established highest-scoring tone. Honorifics, thanks and command-style pronouns are unnecessary in every question; the task, object and division of work matter more. Google's user guide likewise recommends natural language and gives prompts beginning directly with an action. Official guide
References
3. Google, Tips to write prompts for Gemini, official user guide