LOL…My wife gets made at me when I silence an Alexa device by telling it “Alexa, shut up!”
That’s fine (if you no longer wish to discuss this more). If you did someday, though, I’d still be interested in hearing your thoughts on it.
I have a hard time playing evil in videogames, of course I’m nice to AI.
I don’t think any serious thinking person would claim that present-day AI possesses anything like “sentience”, even if they’re not versed in some of the mechanisms of AI. But there are two factors that I think figure into the benefits of being kind or polite to an AI, and both have to do with the level of sophisticated behaviour that they’ve achieved.
The first factor is that AIs act sentient even when they’re not, so being persistently rude to them may shape a pattern of behaviours that carries over into other interactions where it really does matter. Or IOW, for those inclined to act like assholes, it reinforces that behaviour.
The other factor, again related to the very high sophistication of the latest LLMs, is that the nature of a person’s behaviour may affect the way the LLM interacts with them. The effect may be subtle, or it may be as simple as the fact that statements focused on rudeness and insults tend to carry less information than more reasoned ones, hence less constructive dialog ensues.
Thinking about this thread, I opened up both my paid AI’s (Gemini 3.1 Pro and Claude Sonnet 4.6 Adaptive) and said the following:
“Hello, hope you are well!”
Gemini responded with:
And Claude said…
I just recently heard somewhere that the makers / maintainers of AI systems recommend that we are not polite in our interactions. Not that we should be impolite, just state the nature of the request matter-of-factly without any added social niceties. The reason being, the AIs use energy in forming their replies, which is costing the company that maintains the system. Putting a request in polite terms causes the AI to respond politely in kind, using a bit more energy than strictly necessary.
However, having heard of the Roko’s Basilisk thought experiment, I think I will still treat AIs politely…just in case…
I am not “kind” to AI and I typically say “thank you” … just because that is an automatic response in even fictive interactions.
It follows the pattern with my pets. They do not understand what I say to them beyond a smattering of commands and such, but I talk to them anyway. Even say thank you.
Yep, I treat it with the same politeness that I would treat a human. Hey, you never know, so I cover my bases. LOL
Another thought after reading an article in my feed on this -
I wonder if learning to not say please and thank you to a chatbot - for good reasons as above - ends up entraining people to be less automatically polite in real interpersonal casual interactions?
I’m currently listening to Becky Chambers “A Long Way to a Small Angry Planet” and “A Closed and Common Orbit” both of which have AI as main characters, and the tenor of interactions between the ‘natural’ persons and the AI persons drive much of the narrative. A great fun read with some real philosophical depth.
I treat it as something to be flinched back from and avoided. As well as a liar and an agent of people I consider enemies of humanity.
If they were people or even animals I might feel badly about it; as they aren’t, I don’t.