Being Polite to AI Might Influence the Quality of Responses It Gives You… Sort Of

By Danial Martinus  •  04 Mar 2025 in Futurism

Thirsty for JUICE content? Quench your cravings on our Instagram, TikTok and WhatsApp

We’re conditioned as kids to be polite to everyone we talk to. It’s a natural habit that we retain in adulthood, though we probably know when to be our less-polite selves around certain people.

With the use of AI chatbots more rampant than ever before, there comes one question of ethics—should we be polite at all to AI? After all, they’re not even human and are therefore incapable of ‘feeling’ emotions.

According to a joint study by researchers from Waseda University, the RIKEN Center for Advanced Intelligence Project (AIP), and National Institute of Informatics Research and Development Center for Large Language Models (what a mouthful), the way we talk to AI can actually influence the way it responds to us.

AI assistants… or future overlords?

From the ground up, AI chatbots or large language models (LLMs) are built to respond to our prompts in the most concise way possible. The responses usually give you what you need (for the most part) and are easy to tweak with additional prompts.

But when it comes to actually getting quality responses, it’s not a matter of AI understanding that we’re being nice or aggressive to them. Instead, the study finds that AI is designed to mirror our tone or communication style.

For example, if a prompt is curt, the response will be plain and very robot-like. If, on the other hand, we fill our prompts with conversational (human) cues, the response will also sound more human-like, and in a lot of cases, more detailed.

“If a user is aggressive, the bot might try to de-escalate by keeping responses neutral and brief,” says Devansh Agarwal, a machine learning engineer at AWS. “Conversely, in polite exchanges, the bot might offer more detailed answers, since there’s no risk of escalating tension.”

In reality, though, polite prompting doesn’t actually make AI smarter, but it definitely makes it feel more helpful.

But because being polite also means providing more context in a request or prompt, the end result is a more informative and high quality response. These higher quality responses will also encourage users to continue being polite, thus creating even more high quality responses further down the line. It’s a loop of positivity.

If you were to provide a rather blunt prompt, don’t expect anything too helpful. It’s not because the AI is being sassy; it just has very little to work with.

What’s the moral of the story? Being nice and polite goes a long way.

Click here to check out the full study.

For more future-forward stories, head to JUICE Malaysia.