How we talk to AI could shape how we treat people
As AI chatbots continue to integrate even further into our daily lives, some are asking: Does it matter how we speak to them? Last year, OpenAI CEO Sam Altman stated that saying “please” and “thank you” to our AI support pals caused the company to lose millions of dollars in electricity costs. But does being polite to AI actually impact how it operates and the outcomes it provides?
A recent article from ZDNET makes the case that politeness toward AI has very little to do with the machine itself, and everything to do with human behavior. The argument is that repeatedly barking commands at chatbots can normalize impatience and command-driven communication. Over time, those habits may spill into how we interact with coworkers, service staff, or anyone else we communicate with, leading to a world that lacks common courtesy.
However, recent research may complicate things a bit.
What the research says about politeness and accuracy
A very recent study out of Penn State explored whether politeness actually affects AI accuracy. Researchers tested the same questions using prompts that ranged from very polite to openly rude, then measured how well newer AI models performed.
The findings were a little surprising. In some cases, AI models responded more accurately to rude prompts than polite ones. While this doesn’t mean being rude to AI is the way to go, it does show that politeness doesn’t appear to offer a consistent performance advantage.
Where the Spiceworks Community lands on AI politeness
When the Spiceworks Community was asked how polite they are with AI chatbots, they landed in the middle: not overly polite, but not incredibly rude. These answers aren’t surprising considering IT environments are built around efficiency, clarity, and speed and many IT professionals tend to be more direct.
The human side of AI interactions
From a technical standpoint, there’s no strong evidence that politeness improves AI output. However, research does show that AI could shape social norms, expectations, and behavior over time. The concern isn’t that AI replaces human relationships outright, but that AI interactions may subtly change how we relate to real people.
AI systems are designed to be endlessly patient, responsive, and accommodating. They don’t push back, require emotional labor, or ask for reciprocity. Over time, that these cold and frictionless interactions could make human relationships, that are filled with emotions, feel more frustrating. Plus, since AI systems don’t require use to practice empathy or patience, experts are also concerned we may get fewer opportunities to practice patience, perspective-taking, and emotional awareness, which are all key skills for building real relationships.
What this means for AI governance
For IT teams, this isn’t really about teaching people to say “please” to machines. It’s about setting reasonable norms as AI becomes a permanent part of daily workflows. In practice, that can look like:
- prioritizing clear, well-structured prompts over performative niceness
- encouraging professionalism without mandating artificial etiquette
- recognizing that AI is both a productivity tool and part of how teams communicate
- being mindful of tone in shared or collaborative AI use cases
As AI continues to blur the line between tool and collaborator, most IT pros seem to be doing what they’ve always done best: adapting to context, optimizing for the task at hand, and keeping things practical.
And for now, that middle ground might be exactly where it makes the most sense.
What are your thoughts around AI politeness? Let us know on the Spiceworks Community!
