The moral question of our time is: should we say “please” and “thank-you” to artif icial intelligence chatbots? This would be nothing new for me. I routinely talk respectfully to inanimate objects. For example, I always say “thank-you” to my truck (and God) every time it starts.

But AI chatbots are different from trucks or toasters. When we talk to AI, it talks back. A recent study reports that 69% of Gen Zers say “please” and “thank-you” when interacting with ChatGPT. What could be wrong with that? Sam Altman, tech billionaire CEO of OpenAI revealed on X that this politeness is costing him “tens of millions of dollars”. A Washington Post study notes that using AI sucks up a lot of electricity. A 100-word AI email, for example, uses enough juice to power 14 LED lights for an hour. Sending one of these every day for a year would add up to 7.5 kWh, or enough electricity to power 9 entire households for an hour. Therefore, including any absolutely unnecessary words can be costly. Also, cooling off these big AI data centers consumes millions of gallons of water in this era of world-wide water shortages.

Being polite to AI raises other questions. And who better to ask if we humans should be polite to AI than AI itself? ChatGPT began its answer to my question by praising the question as thoughtful and interesting—which made me instantly suspicious. It replied, “AI like me doesn’t have feelings, so you won’t hurt or offend us by being rude.” “Like me?” Was ChatGPT now implying it had a separate identity? And by referring to AI as “us” wasn’t it suggesting that humans were “them”, the other, drawing up sides in a potential conflict?

ChatGPT went on to say that being polite and respectful to AI reinforces good habits in our interactions with other people and being rude to robots normalizes bad social behavior. Okay, that makes sense. Apparently, the golden rule of “do unto others” also applies to AI. A Microsoft WorkLab memo says of AI, “When it clocks politeness, it’s more likely to be polite back.” But is AI training us to have good manners or does it have ulterior motives?

Sure, it is great to have a little AI buddy who knows everything and appears eager to answer all our questions without making us feel stupid. But are we confident that AI is really working for us? While we are gathering information from AI, it’s gathering information from us, creating a database of our interactions, both polite and impolite.

Is AI a cyber-Santa Claus who knows if we’ve been naughty or nice? And if we’ve been naughty, couldn’t AI get revenge by giving us faulty medical advice or pushing us to buy stock in Tesla? Will AI pass on our profiles to future generations of household robots who “remember” our previous disrespect? Will our robots be like J.D. Vance and when we least expect it, will they turn on us with glowing red eyes and say, “Did you even say ‘thank-you’?”