The Guardian asks whether you should say 'thank you' to Alexa and other AI assistants

The Guardian has published a selection of reader responses to a question that has long since stopped being just a household habit: should you be polite to voice assistants and AI services? The discussion quickly moved away from the feelings of machines and toward how these conversations change the people having them.

Where politeness comes from

Some participants are convinced that saying "please" and "thank you" to Alexa or a chatbot benefits the user, not the device. In this view, politeness works like muscle memory: people repeat the same communication pattern over and over, then carry it into everyday life. This matters especially for children and teenagers, who increasingly talk to technology at least as much as they talk to people. If a curt, commanding tone becomes the norm at home, it can easily seep into conversations with cashiers, colleagues, or family. One reader put it as simply as possible: even if AI feels nothing, people are still responsible for their own manners.

Several responses returned to the same idea: rudeness toward an interface that can't answer back is convenient precisely because it carries no social cost. But the habit of speaking harshly, venting frustration, and barking orders without softening them gets reinforced over time. For supporters of this view, being polite to a machine is a form of self-discipline, not an act of sympathy.

"They're not people, but I am a person — and I try to be a decent one."

The case against

Skeptics in the same discussion point out that excessive politeness toward AI has side effects of its own. Their main argument is that we shouldn't humanize systems that only imitate engagement, understanding, and emotion. If you keep responding to them as if they were a conversation partner with an inner life, the line between a tool and a quasi-person starts to blur. For some readers, this is not a harmless verbal habit but a cultural shift: people increasingly attribute intentions, character, and even moral status to a machine that has none of them.

There's also a more down-to-earth argument: extra words in every request mean extra computation. One participant directly linked "thank you" and "please" to additional energy and water consumption, while another recalled Sam Altman's remark that responding to users' polite phrases costs OpenAI tens of millions of dollars. From a practical standpoint, the argument goes like this: if AI is a utility, why load it with phrases that don't change the meaning of the command?

Where to draw the line

In the end, the discussion went beyond a simple "yes" or "no." Even many supporters of politeness clarify that they don't consider assistants living beings and regularly remind themselves of that. Their position is more about everyday speech hygiene than about recognizing AI as a subject. Skeptics, for their part, aren't necessarily calling for rudeness: they suggest speaking briefly, clearly, and without theatrical courtesy. The reader responses show that the real question is broader: what should our standard be for talking to machines that speak in a human voice?

• Politeness as practice for everyday manners
• The risk of carrying a commanding tone into conversations with people
• The danger of humanizing a soulless tool
• Extra energy, water, and compute costs

What it means

The debate over saying "thank you" to Alexa turned out to be a marker of a bigger issue: interfaces are becoming so conversational that people have to renegotiate the rules for talking to technology. While some see politeness as a useful social reflex, others want to keep a hard line between human and machine. For the industry, this is an important signal: the more human AI sounds, the more a product affects not just convenience but users' everyday habits.