Key takeaways of zdnet
- Openai has made GPT-5 warmer and friendly.
- Many users say they do not want a friendly chat.
- There is a challenge in finding a balance between simplicity and choice.
Get ZDNET Tech coverage more deeply: Add us as a favorite google source Chrome and chromium on the browser.
Looking for a Kinder, Gentler AI model? Then you can prefer the latest taste of GPT-5. Other people, however, not so much. On Friday, Openai announced an update for its latest AI model, designed to improve its personality and calm his critics.
Too: What is the price of $ 20 still the chat plus provides so much – including GPT -5?
“We are making GPT-5 warmer and friendly on the basis of feedback which also formally formally,” OpenaiI Said on x“Changes are subtle, but Chatgpt should now feel more acceptable. You will look at small, not small, real touch, such as ‘good questions’ or ‘great beginnings’. Internal tests show no increase in sycophancy than the previous GPT-5 personality. Changes may be to roll out for a day, more updated.”
(Disclosure: ZDNET’s original company Ziff Davis filed a case of April 2025 against Openai, alleging that it violates Ziff Davis copyright training and operating its AI system.)
Reactions so far
Well, it seems like a meaningful change, one is to give AI a more sociable and friendly personality, without being much more clear about it. But so far based on the user’s response, it is not necessary what people want.
An X user said in response, “It’s not at all what people want. I do not want it to pretend to be human, I want it to spread this information parsimonically.”
“Don’t talk to me to imitate a human. You are a machine, work like it,” another said.
“I don’t want it, I want it to be my help, not my friend.”
“Please don’t put it on Pro. I want I answer my questions. If I want the real human conversation I go out instead,” a fourth added.
Too: Don’t like GPT-5? You can still use GPT -4 and other heritage models in chat – how is here
There were reactions of some people mixed with complaints who preferred changes. Others suggested offering “personality” as a togal, leaving each user to decide how AI should work. In general, however, most want to directly answer their requests without any disturbance or frills.
Ongoing complaints
Since the onset of GPT-5 earlier this month, Openai has been stunned by a wave of protests on the new AI. Some users who were hypnotized about the release of GPT-5 have found it a lethargy. Others have reduced the sudden loss of GPT-4 and consider less than competent compared to GPT-5.
To address the ongoing complaints, Openai has twice-5 and twenty-five. Trying to calm old models prioritizing people, the company restored heritage GPT-4o, GPT-4.1, O4-Min and O3 models, but only for plus and pro customers.
Too: I tested the GPT-5 coding skills, and it was so bad that I am sticking to GPT-4o (for now)
The GPT-5 was also considered a more efficient model that would decide about how the best to answer your request, choose a quick answer, long and deep thinking, or any other difference. But people folded on that change, openi added four different modes – for those who do not want to answer or AI needs to tell how to answer, think fast for quick answer, to think for a long time, and think of mini mode for sharp thinking.
Rocky reception
The rocky reception that greeted the GPT-5 raises questions about what we want in an AI and it can go as everything for all people. Do we want a simple cut-to-chase robotic assistant who can answer our questions quickly and in summary as much? Or do we want a friendly voice that can provide warmth and etiquette with reactions? In many cases, it depends on the request.
If I am asking AI to answer a factual question, generate code, or provide deep scholarly research, I want it to be quick and efficient without wasting time on pleasures. But if I am taking advice on a personal matter or need a sounding board, I would like some personality to separate with response. Chatgpt even allows you to customize it so that you can keep an eye on individual details that you want to share.
Too: All GPT-5 updates here are Openai rolled out since launch
With GPT-5, Openai tried to make the chatgpt simple and more efficient, ending the giant array of models and mode. Now, the company is going back in the opposite direction by bringing back all different options. Since AIs like chat can not be all things for all people, the challenge is to balance between simplicity and user’s choice.
It may be that the twix further reaches the right place.
