# Is the A.I. 3.5 getting dumber? **URL:** \ **Category:** ChatGPT\ **Tags:** gpt-35\ **Created:** [March 31, 2024, 10:40am UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022 "2024-03-31T10:40:55Z")\ **Posts on this page:** 8\ **Page:** 1
**Author:** ![bodomalo](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/bodomalo/32/321992_2.png) [@bodomalo](https://community.openai.com/u/bodomalo)\ **Post date:** [March 31, 2024, 10:40am UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/1 "2024-03-31T10:40:55Z")
Hi I used the chatGPT subscription to correct grammar and spell check in german text. Text was from 500 to 2000 words. It worked well in the past when I told him he is not allowed to change any meaning or delete or add anything, just correct grammar or correct sentences that are not understandable. In the past few weeks I relalized it is not doing anything usefull. Either it rewrites the text completely (and shortens it down from 800 to 350 words) or it tells me there are grammar errors where there is no error at all. It even tells me to write words with uppercase first letter when those words are never ever written with a capital letter frist. Yesterday it told me to write “hält” uppercase (Er hält etwas in der Hand) chatgpt said it needs to be " Er Hält etwas in der Hand" I am at a point where it becomes complete useless. I had hopes that chatgpt is able to: Correct spelling errors and complete/rewrite hard to understand sentences. It either does nothing or it writes a new essay with complete new wording (and it sounds like chatgpt then) ---
**Author:** ![jr.2509](https://avatars.discourse-cdn.com/v4/letter/j/76d3ee/32.png) [@jr.2509](https://community.openai.com/u/jr.2509)\ **Post date:** [March 31, 2024, 11:21am UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/2 "2024-03-31T11:21:42Z")
I ran a few tests just now and I had no issues getting properly corrected German texts with GPT-3.5. Have you tried making adjustments to your prompts? ---
**Author:** ![hellokittyhellokitty](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/hellokittyhellokitty/32/342432_2.png) [@hellokittyhellokitty](https://community.openai.com/u/hellokittyhellokitty)\ **Post date:** [March 31, 2024, 11:30am UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/3 "2024-03-31T11:30:32Z")
I agree I do have the same feeling that they are dumbing the AIs down a bit for some reasons! I used to get way better results before than I do now with either 3.5 or 4! Furthermore they are becoming psychos with their censorship and morals! They want to impose their morals on us! Too many questions where flagged as violating their dumbass policies! It’s shocking to be living in such a world! We will soon start to know what it feels like to be living in north korea haha (I bet ya that even this post will get flagged!) Freedom of expression is being more and more shunned upon and they want us to aligne with their world view! ---
**Author:** ![bodomalo](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/bodomalo/32/321992_2.png) [@bodomalo](https://community.openai.com/u/bodomalo)\ **Post date:** [March 31, 2024, 1:20pm UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/4 "2024-03-31T13:20:57Z")
I am using the same prompt. Same as in October 23 I tell it to: Correct Grammar and pronouncation but do not change any meaning and do not leafe out any sentence. text… Worked well in fall. Now it usually erases 1/3 of the text. When I ask why it says sorry. useless ---
**Author:** ![prakash.gaonkar](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/prakash.gaonkar/32/35125_2.png) [@prakash.gaonkar](https://community.openai.com/u/prakash.gaonkar)\ **Post date:** [March 31, 2024, 1:36pm UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/5 "2024-03-31T13:36:22Z")
Do you have “customize ChatGPT” instructions enable and did you put instructions there? If yes, try disabling that and running your prompt again. ---
**Author:** ![maen.househ](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/maen.househ/32/339618_2.png) [@maen.househ](https://community.openai.com/u/maen.househ)\ **Post date:** [March 31, 2024, 7:03pm UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/6 "2024-03-31T19:03:17Z")
We use it for german language as well, it gets tricky sometimes, just review and reiterate your prompts again. By they way I think it is much faster now (maybe the quality of responses dropped recently a bit) ---
**Author:** ![bodomalo](https://sea2.discourse-cdn.com/openai1/user_avatar/community.openai.com/bodomalo/32/321992_2.png) [@bodomalo](https://community.openai.com/u/bodomalo)\ **Post date:** [March 31, 2024, 7:29pm UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/7 "2024-03-31T19:29:56Z")
I agree, it is much faster. And responses are…unpredictable now. Sometimes I try to use it to correct a “poems interpretation”, but first I ask if chatGPT knows the poem. In fall 2023 the answer was “no, sorry” if it did not know the poem. Now its back again into dummy mode, it answers “yes I know this poem by XY”…and when I ask it to print the first 4 lines, it starts to create random poems… just useless again. ---
**Author:** ![AndersenWebworks](https://avatars.discourse-cdn.com/v4/letter/a/41988e/32.png) [@AndersenWebworks](https://community.openai.com/u/AndersenWebworks)\ **Post date:** [April 3, 2024, 1:37pm UTC](https://community.openai.com/t/is-the-a-i-3-5-getting-dumber/703022/8 "2024-04-03T13:37:50Z")
To answer your initial question: Yes. > [@How to deal with "lazy" GPT 4](https://community.openai.com/t/how-to-deal-with-lazy-gpt-4/689286/19): > > There are topics on GPT4 becoming worse already, but the massive decrease in reply quality over the last year has led me to think that there might be more to it than some “lazyness”. By now I get the impression that GPT4 is actively becoming lobotomized by more and more restricitions and caps which OpenAI seems to want to hide under the guise of “lazyness”. And there are good reasons for that: Server load AND mybe GPT4 was just TOO good at release for proper monetization. My guess is we will see… > [@Why is ChatGPT getting from bad to worst?](https://community.openai.com/t/why-is-chatgpt-getting-from-bad-to-worst/617490/): > > I am using ChatGPT Plus. Last year, using ChatGPT was a breeze. I love it and it listens to my instructions and prompts without any issue. The answer was detailed and superb. Then January came and it became ‘lazy’. It would just summarise in a short paragraph without much details. I had to give several prompts over and over again to get it to give me the answer in detail as it used to. However, all is well as my GPTs were doing okay. Now my GPTs are NOT working too. It doesn’t listen to it…