According to a recent Delve Research / Invisio national pulse survey 62% of Irish users are – and this can have some benefits.
On our recent national pulse survey of 565 people1, 62% of AI users said they were polite with their AI tools, using words like “please” and “thank you” in their prompts.
Comparable figures observed elsewhere range from 58% to 67% in the U.S. and 71% in the UK. We will continue tracking the Irish figure on future waves of our pulse survey.
Professional in = Professional out
With LLMs like ChatGPT and CoPilot, what you put in often shapes what you get out. A member of the design team for MS CoPilot reminds us that the language we use in our prompts sets the tone for the subsequent conversation, so that using a polite and professional tone in your prompt will generate responses pitched at a similar level.
38% of our respondents said that they were “not polite but to the point, I ask and expect an answer”. In the context of setting the tone for the subsequent output, this can also work well especially if that tone aligns with your expectations of the interaction.
Then there’s the 2% who said they were “sometimes impolite—I’ll swear or be abrupt with AI.” This may be one of the few safe spaces left to blow off steam—unless you believe that our future AI overlords are taking notes. In which case... good luck.
Beyond tone-matching or future robot retribution, there's also a human angle: being polite to your AI may help reinforce your own sense of humanity in an increasingly automated world.
Are you polite to your AI? Do you sometimes vent your frustration to your robotic assistant? We’d love to hear your experience!
If you would like more detail on our survey results, or detail on how to include questions on our next national pulse survey, please get in touch with us today.
1 Margin of error +/- 4%
Blog feature image generated using Sora