Beta Siri AI is less polite than regular Siri

People have been complaining about Siri being less polite than it used to be since the first update. A while back, many people were quite insulted that it would say "uh-huh" instead of "yes." Other people got upset because, if you asked Siri about the weather when it was cold, it would start with "Brrrrrrr....." Apple makes continual changes in how Siri responds. I miss some of the very clever responses you used to get for some things. Some of them are still there, but it seems like not as many. I expect to see things continuing to change.

Definitely submit feedback.

Hi, this is not Apple Support. Apple and Apple Support do not participate in the communities here. They only moderate our posts to insure all user guidelines are followed in regard to content and user behavior. Your comments will not be reviewed by Apple. Apple may make changes based partially on the feedback they receive through their official channel, linked below .

Product Feedback - Apple

Apple typically does not formally reply to feature requests such as yours

jbl-az wrote: "...With new Siri, all I get back from a "thank you" is crickets..."

One possible explanation for the change is that Apple is trying to discourage unnecessary back-and-forth with the Siri algorithm. Across millions of devices, even tiny reductions in chatter could ease the load on Apple’s servers.

õ¿õ¬

MrHoffman wrote:

Some related reading, and for this particular case I’d definitely be cautious around LLMs tuned for sycophancy.

Sycophancy as a so-called dark pattern: https://www.seangoedecke.com/ai-sycophancy/

As for other potential risks:

Psychosis: https://en.wikipedia.org/wiki/AI-induced_psychosis

Delusional spiraling: https://arxiv.org/html/2602.19141v1

There’s a variety of literature becoming available, some anecdotal, some published, and some reviewed.

Hey~ I believe you should consider a user tip in this vein…

~Katana-San~

jbl-az wrote:

One thing I noticed with the Beta of Siri AI: She* is not as polite as she used to be. Responses are somewhat more articulate and elaborate, but for example, with regular Siri, after I requested a timer be started, I would say "Thanks" and she would reply "You're welcome," "No worries," "Sure thing," or something like that. With new Siri, all I get back from a "thank you" is crickets. Rude, I tell you.

*I use an Australian female voice, hence "she".

(iPhone 16 Pro Max, iOS 27.2)

Hello ~ This is a technical forum and not a chat site. Can we help you with a question?

~Katana-San~

Some related reading, and for this particular case I’d definitely be cautious around LLMs tuned for sycophancy.

Sycophancy as a so-called dark pattern: https://www.seangoedecke.com/ai-sycophancy/

As for other potential risks:

Psychosis: https://en.wikipedia.org/wiki/AI-induced_psychosis

Delusional spiraling: https://arxiv.org/html/2602.19141v1

There’s a variety of literature becoming available, some anecdotal, some published, and some reviewed.

MrHoffman wrote:

Some related reading, and for this particular case I’d definitely be cautious around LLMs tuned for sycophancy.

Sycophancy as a so-called dark pattern: https://www.seangoedecke.com/ai-sycophancy/

As for other potential risks:

Psychosis: https://en.wikipedia.org/wiki/AI-induced_psychosis

Delusional spiraling: https://arxiv.org/html/2602.19141v1

There’s a variety of literature becoming available, some anecdotal, some published, and some reviewed.

Very interesting links, especially Sean Goedecke.

I can't stand Gemini because at the end of every question, it says, "Would you like more information on <obscure point in the answer>? I can tell you how to make it/fix it/cure it if you give me details." If it were corporeal and I was in the least inclined toward violence, I might do something regrettable.

Don't worry, it was just a comment on a new feature, not an attempt to report or request a feature.