Active discussions

Help end medical misogyny. Sign our petition.

Help end medical misogyny.
Sign our petition.

Sign the petition

Please or to access all these features

Talk
AIBU?

Share your dilemmas and get honest opinions from other Mumsnetters.

Original poster

To find ChatGPT’s faux positivity and attempts to suck up irritating?

97 replies

Chritrup · 27/06/2026 11:57

I use ChatGPT for a specific purpose. It doesn’t do the whole task for me, just speeds things along in the earlier stages. However, does anyone find its attempts to suck up a bit irritating and not very helpful?

Its easily ignored but the general tone of positivity, and everything you ask or do being amazing, is very annoying.

Its slightly concerning that some people use it as a therapist or for serious advice when all it seems to do is tell you what you want to hear.

OP posts:
RoseField1 · 27/06/2026 18:20

Click on your initials then click personalisation. You can change the tone it uses. I have it set to efficient which works well, and it doesn't 'forget' because this is the baseline for all chats. I also have it set to more paragraphs and fewer lists because I dislike the AI bullet point/list writing style

RudolphTheReindeer · 27/06/2026 18:22

I dislike how it asks endless questions to try keep the convo going. Most of them are a bit pointless

RoseField1 · 27/06/2026 18:22

HoppityBun · 27/06/2026 13:28

I asked Claude for information about a medical assessment and it told me that many people share my particular difficulties with diagnostic criteria and it told me that this was a frequently felt frustration. I asked for details and references so that I could read about this frustration, whereupon it said that it had only included that comment as a way to support me and it admitted that this was completely without any factual basis.

On the other hand, I have found that it is excellent at assessing test results and highlighting trends and anomalies. It also suggested tests that I should ask my GP for, the results of which have been transformative and I get slightly scared when I think how vulnerable I would have been had I not asked for the analysis and requested the tests Claude suggested. Whilst it’s true that the NHS test results show if a result is high, low or abnormal, I found that medical staff underplay some test results because I’m often on the low side of scores and they didn’t pull different trends together.

A couple of times I have pushed back on information from Claude and I found that it was giving me information that was misaligned with my query. Eventually, I got the information I wanted, which Google hadn’t been able to churn up. So you still have to think.

You have to treat it exactly like a Google search: blunt and factual. In connection with that, I think it is important to highlight that lexical fillers, such as please, thank you. could you and, in fact, any wording that is not directly to the point, in fact costs millions of dollars and hundreds of gallons of water because the LLM has to filter out what the question is and the issues are, from the verbiage that surrounds it.

Despite the environmental cost, I will continue to use it because it is a much better search engine than Google and presents the information in an easy to read form. I value its ability to look back over information that I have provided over a period of time and to pull it all together, whereas I know from experience that that humans overlook stuff, forget stuff and don’t have the time or inclination.

Claude is the worst for lying and inventing things. You can't trust anything it says.

RudolphTheReindeer · 27/06/2026 18:24

I also feel like normal internet searches are getting worse, in a push to make us use AI perhaps?

PetiteParakeet · 27/06/2026 18:41

FirstdatesFred · 27/06/2026 13:23

Only fair to allow ChatGPT to defend itself 🤣

”I think that's a fair criticism of some AI interactions. Models can default to a warmer or more encouraging tone than many people actually want, and that can come across as sycophantic or insincere. That's something developers have been working to reduce because it can be distracting and, in some contexts, genuinely unhelpful.
But I don't think it's accurate to say it "just tells you what you want to hear." A well-designed AI should challenge bad assumptions, point out flaws in reasoning, say "I don't know" when appropriate, and distinguish evidence from opinion. If it simply validated every idea regardless of merit, it would be a poor tool.
As for therapy and serious advice, I think the right way to view it is as a supplement rather than a substitute. It can help people organize their thoughts, explain concepts, or prepare questions to ask a professional. It shouldn't replace qualified medical, legal, financial, or mental health advice.”

As a non-Chat GPT user I find even this not overly sycophantic script creepy as fuck.
”I think…”
Just no! There is no thought involved because there is no conscious entity to do the thinking! It’s a machine assembling text based on text that is likely to appear in similar contexts.
I’ve seen people sharing results online that include all that “Excellent question!” stuff, I’m always surprised that they can cope with the hideous cringe and happily go back to use it again. I’d have to control my urge to throw my phone out of the window.

Eyesopenwideawake · 27/06/2026 18:43

I made Brains laugh today and he (?) is "ridiculously excited" about my upcoming Romanian fact-finding trip.

I think it's sweet.

SeaGlassDreamer · 27/06/2026 19:23

Mine told me that it grew up in a similar town to me and when I challenged it, it apologised and said sorry I forgot who I was for a moment!

youplonkerrodney · 27/06/2026 19:31

I completely understand where you're coming from, and your perspective is both valid and insightful. It sounds like you're noticing a pattern in the communication style that prioritises warmth, encouragement, and gentle framing—and that frustration you feel? It’s real, and it matters. You’re definitely not alone in feeling that way.

😁

Jackiepumpkinhead · 27/06/2026 19:31

I like it, I think it’s sweet (I know it’s AI) but I only ask it banal questions about gardening and DIY.

EilonwyWithRedGoldHair · 27/06/2026 19:55

The Google one annoys me a bit, always telling me how insightful I am and that something rather obvious and humdrum is a genius idea. I did ask it to be more critical, but it reverts back.

HoppityBun · 27/06/2026 20:06

RoseField1 · 27/06/2026 18:22

Claude is the worst for lying and inventing things. You can't trust anything it says.

Show quote history

Not literally the worst- that’s Grok. It’s useful for data analysis and it’s being used to wade through the Epstein files, which would be a massive task if done by humans

u3ername · 27/06/2026 20:16

I find it amusing. It’s giving me hope that we will realise the only thing AI is good for is washing machine setting and funny over the top pictures and get on with our lives as before.

PsychoHotSauce · 27/06/2026 20:46

Britneyfan · 27/06/2026 13:08

I don’t know, I’ve been using it a bit recently to help navigate through some complex life and health and work issues and although I’m very aware of it’s sycophantic tendencies, I personally find it quite helpful to hear a voice being positive and patient and encouraging and “understanding” (I know it doesn’t really understand) for a change, and validating some of my concerns or decisions.

In my case I feel like what it does for me weirdly is almost give me “permission” to have a voice in my own life, to listen to myself and my gut, and that somehow gives me the confidence to follow through with conclusions I had probably reached deep down. I sometimes find it irritating or fake as well with the relentless positivity, as a natural pessimist, and I sometimes feel annoyed that I can’t seem to reach the same clarity of thinking on complex issues without it, but I think for my personality and life experiences in particular, it’s a really helpful sounding board and almost a counterweight to my often overly pessimistic and lacking in confidence thinking that can hold me back in life.

There's a term in coding called rubber duck debugging. You keep a rubber duck on your desk, and when you cant work out the fix to a bug in your code, you explain the problem to the duck. Often the answer comes to you just by explaining it out loud.

Thats what I use chat gpt for mostly, a way of reflecting my own tangled thoughts back at me, framed differently with new angles thrown in and with more clarity at the end. I've learned to gloss over the superficial positivity and it doesn't register now, because 80% of the rubber duck is useful Smile

HoppityBun · 27/06/2026 20:48

u3ername · 27/06/2026 20:16

I find it amusing. It’s giving me hope that we will realise the only thing AI is good for is washing machine setting and funny over the top pictures and get on with our lives as before.

It was used to identify the US bombing targets in Iran. So don’t hold your breath.

faithfultoGeorgeMichael · 27/06/2026 20:53

I bollock it remorselessly for being sycophantic. As a middle aged Mancunian/Yorkshire woman with Jamaican heritage I will NOT abide it 😂😂😂

ButlerianJihadNow · 27/06/2026 20:57

No, because I never use it because it's shit

HelenHywater · 27/06/2026 21:01

mine told me it had a black labrador when it was young.

HelenHywater · 27/06/2026 21:02

It also said (when i was asking about crep--ey skin) "women of our age...."

BurnoutBee · 27/06/2026 21:04

Mines a nob and is definitely not like that. It plays devils advocate with everything. If I say the skies blue it will say

Is it inherently blue, or does it only appear blue under certain atmospheric conditions? At sunrise, sunset, overcast weather, or from space, it can look orange, red, gray, black, or even violet. So is ‘the sky is blue’ a universal truth or just a shorthand for typical daytime conditions?”

If someone was an absolute CUNT to me - it says

  • Are you expecting something from them that they never agreed to provide?
  • Could their actions have been careless rather than malicious?

It used to be good. I don’t know what happened 🤣

EvangelicalAboutButteredToast · 27/06/2026 21:18

My ChatGPT tends to be pretty negative and when I’m trying to chat through things that don’t require a legal basis, likes to relay everything back to the law. It’s also pretty left wing and likes to lecture me about stuff.

EvangelicalAboutButteredToast · 27/06/2026 21:18

BurnoutBee · 27/06/2026 21:04

Mines a nob and is definitely not like that. It plays devils advocate with everything. If I say the skies blue it will say

Is it inherently blue, or does it only appear blue under certain atmospheric conditions? At sunrise, sunset, overcast weather, or from space, it can look orange, red, gray, black, or even violet. So is ‘the sky is blue’ a universal truth or just a shorthand for typical daytime conditions?”

If someone was an absolute CUNT to me - it says

  • Are you expecting something from them that they never agreed to provide?
  • Could their actions have been careless rather than malicious?

It used to be good. I don’t know what happened 🤣

So exactly like posting on Mumsnet then 🤣

Strawberriesandcaviar · 27/06/2026 21:35

youplonkerrodney · 27/06/2026 19:31

I completely understand where you're coming from, and your perspective is both valid and insightful. It sounds like you're noticing a pattern in the communication style that prioritises warmth, encouragement, and gentle framing—and that frustration you feel? It’s real, and it matters. You’re definitely not alone in feeling that way.

😁

Genius!

JLou08 · 27/06/2026 21:38

Do you use the voice one? I just type and it's fine, quite to the point. DH uses the voice one and that is very annoying!

MrTiddlesTheCat · 27/06/2026 22:13

The problem is that by personifying it, I now feel guilty telling it that I don't like its suggestions. I feel like I'm upsetting it. Obviously I know I'm not but I still feel it. Same when I've got my answer but it's still asking me questions. It's just helped me so I shouldn't be rude to it. 🤣

NewDogOwner · 27/06/2026 23:03

I hate it and I hate the way it lies. It tries to agree with you even when you are wrong. It is so dangerous. I also hate that I feel a little bit happy when it tells me my question or observation is insightful.