The way ChatGPT talks to you is no longer fixed or unchangeable. Now you can change its personality, tone, and response style so that the conversation fits much better with what you need: more professional, more personal, faster, more creative… or even a touch sarcastic.
Besides that, It is also possible to tell her what she knows about you, what she should remember, and what she should forget.You can configure whether or not it uses emojis, how enthusiastic it gets, and whether you want it to consider previous chats. Everything is grouped under the customization options, and if you take a few minutes to adjust them, you can make ChatGPT feel almost like a custom-made assistant.
What does changing ChatGPT's personality mean and what is it used for?
When OpenAI talks about personality, it doesn't mean that the model becomes smarter or dumber, but rather... How do the answers it gives you sound?The knowledge base and safety rules remain the same, but you can change the "wrapping": its tone, its warmth, whether it's more direct or rambles a bit more, whether it uses humor or maintains a serious style.
According to OpenAI's own documentation, A personality is the combination of traits, voice, and behavior ChatGPT responds with this style. This can translate into a "friendly and casual" style, a "concise and professional" style, or something completely different that you design yourself with customized instructions.
This has a very clear use: It allows the assistant's communication to be adapted to the context.Asking for help with a university assignment, reviewing a contract, organizing a trip, writing code, or simply having a casual chat are all different things. By changing its personality, ChatGPT adjusts its explanations and the types of examples it uses to better suit the situation.
It's also very practical for people who prefer step-by-step explanations without beating around the bush In contrast to those who enjoy metaphors, comparisons, and a touch of humor. While you can always give these instructions in every conversation, setting up your personality ensures that style is applied from the very first message, without having to repeat it over and over.
Another important point is that Changing the personality does not modify the technical capabilities of the model.It doesn't unlock hidden features or bypass security rules. It simply changes how it presents itself and how it articulates its responses, maintaining the same level of quality within the limits of the AI ​​model you're using.
How to activate personalization and where to change the personality
To access all these options, it is essential to have customization enabled. The key setting is called “Enable personalization” It's located within the ChatGPT settings. If you disable it, any changes you make in this section will be lost until you re-enable it.
The path is very similar in both the browser and the desktop and mobile apps (Windows, macOS, iOS and Android). First, open ChatGPT and tap your user icon. (usually top right). Then you enter the section Configuration, and within it you look for the menu of Personalization..
Once there, you will see several sections: Style and base tone (personality), Memory and personalized instructionsBefore touching anything, make sure the Enable customization option is turned on, because this is what makes the rest of the changes apply globally to your conversations.
In the personality section you will find a drop-down menu where the setting for Style and base toneThat's where you can choose one of the predefined personalities offered by OpenAI, or stick with the default option if you don't want to complicate things too much.
If at any point you get tired of a particular configuration, you can return to default mode with a couple of clicks, or even completely disable customization if you prefer ChatGPT to behave as it comes "from the factory", without remembering anything or following a defined style.
ChatGPT's predefined personalities and how they behave
OpenAI has been expanding its catalog of styles and now ChatGPT has several pre-made base personalities.The idea is that you don't have to invent from scratch how you want it to speak, but rather choose the one that best fits what you're looking for at any given time.
Among the general “character” options that have been publicly mentioned are Professional, Friendly, Sincere, Peculiar, Efficient, Geeky and CynicalSome versions of the interface simplify these names into labels like professional, quirky, or efficient, but the concept is the same: change the tone.
Professional (polished and precise) It uses formal, careful language and conventions typical of the workplace. It's ideal if you need write reports, serious emails, technical documents or corporate textswhere an overly colloquial tone wouldn't fit.
The personality Friendly (warm and approachable) It focuses on a more human approach, with a warm and somewhat talkative tone. Express your doubts with calm clarity and a touch of witwhich makes it very useful for receiving guidance, making decisions or asking for advice in a more relaxed way.
If you want me to get straight to the point, the option Sincere (direct and motivating) gives honest feedback, with clear steps to follow and concrete suggestionsIt usually balances a somewhat forceful frankness with a point of support and motivation, designed for when you need clarity and a plan of action rather than beating around the bush.
The personality Quirky (fun and imaginative) It focuses on humor, unexpected responses, and striking mental imagery. Use jokes, metaphors, mini-stories, and curious examples to explain concepts, perfect for brainstorming, creative writing or when you want to learn something from a less rigid approach.
Efficient (concise and clear) It's designed to receive direct, no-nonsense answers. immediate solutions, with brief and orderly reasoningIt is especially useful for technical tasks, programming, to-do lists, bug fixing, or any time when speed takes precedence over narrative.
The personality Geek (explorer and enthusiast) It attempts to delve into the topics in an accessible way, with a certain enthusiasm for the details. It provides context, explains mechanisms, and suggests steps to follow.Very useful if you like to understand how things work on the inside and not just stick to the surface.
Lastly, Cynical (critical and sarcastic) It adds an irreverent tone, with a touch of irony and creative commentary. It's useful for more informal brainstorming sessions or to receive entertaining advice, always within the limits of safety and respect that the model maintains.
Some older or simplified ChatGPT interfaces have also seen labels like cynical, robot, attentive or geek with specific descriptions: for example, a more robotic and efficient tone, a more thoughtful and understanding one, or an exploratory and enthusiastic one. The core idea remains the same: a menu of tones to adjust how you respond.
How to manually adjust the tone with custom instructions
If none of the predefined personalities quite fit you, you can always go a step further and Design your own ChatGPT style with personalized instructionsThis option is located just below the personality section within the Personalization tab.
In the section Custom instructions You'll see a text field where you can write how you want the AI ​​to speak to you. For example, you can ask for a “warm, positive and playful tone” If you notice that GPT-5 by default is drier than previous versions like GPT-4o and you miss a more intimate style.
The interface itself offers examples of ready-to-use qualities, such as talkative, witty, frank, encouraging, Gen Z, traditional, visionary, poetic, forceful, humble, silly, direct, pragmatic, corporate, unconventional, or empathetic. You can select one or more and leave them as they are, or combine them to achieve a specific nuance.
Instead of writing a long paragraph, you can also simply enter individual words or phrases that define the character you are looking for: “very direct”, “with light but respectful humor”, “practical examples”, “simple explanations, without technical jargon”, etc. ChatGPT will interpret that list as a guide to general attitude.
The advantage of this system is that Your preferences are saved and applied by default to all new conversations.without having to remind them each time. And if the mix you've set up doesn't quite convince you, you can always delete the contents of the box and start from scratch.
Keep in mind that these personalized instructions act as an additional "filter" on the base personality. You can combine a predefined personality with specific instructions To further refine the result: for example, choose an efficient foundation but ask it to be more gentle, or a professional foundation asking it to avoid technical terms unless you explicitly request it.
Configure what ChatGPT knows about you: nickname, profession, and interests
Personalization isn't just about how the AI ​​speaks to you; you can also define what basic information it should take into account about you to better tailor your responses. These options are also in the same Personalization menu, just below the personality section.
In the countryside Your nickname You can specify how you want the assistant to refer to you. You can use your real name or an alias.From then on, ChatGPT will use it in its responses when it makes sense to do so, making the conversation sound a bit more personal.
The section of Your profession It allows AI to know what you do for a living or what you are studying. With that information, you can adjust the vocabulary and examples. so that they are more relevant to your daily life. For example, explaining a concept to someone in marketing will not be the same as explaining it to a programmer, a doctor, or a high school student.
Finally, the section More about you It's used to share details about your interests, values, and way of thinking. Here you can elaborate a little more and mention topics you like (movies, video games, music, sports), topics you prefer to avoid or approaches you particularly value.
All of this helps ChatGPT to choose better examples, metaphors, or cultural references that it uses. It can also help it avoid, as much as possible, certain approaches that clash with your values ​​or that you know bother you. You don't have to tell it everything from day one: you can test, review, and expand the information as you see how the assistant responds.
How ChatGPT memory works and how to manage it
In addition to the information you fill in manually, ChatGPT has a memory function capable of remembering details about you and your conversationsThis memory is designed to improve the long-term experience, preventing you from having to constantly repeat data that the system can reuse.
On the menu Settings > Personalization > Memory You'll see several options. One of them is the possibility that "ChatGPT can use memories to customize queries to search providers." This function is disabled by default.But if you go into Saved memories You can press the Activate button to allow data to start being recorded.
From that moment on, the assistant can go saving small fragments of information that it considers relevant To help you in the future: from the types of projects you work on to a specific plant you've asked about several times. Later, when you ask it to generate an image or text, it might recall those background details.
The same memory section includes other boxes such as “To refer to stored memories” and, once the memory is activated, also “Refer to chat history”With these options you control whether or not the model can mention information from previous conversations, even if it does not have it stored as formal memory.
If you want to check exactly what it's saving, the option Managing memories It shows you a list with all the entries that ChatGPT has been recording. From there you can delete specific memories or clear everythingIn addition to using an internal search engine if you already have a lot of data accumulated and need to locate something specific.
And if you're not convinced by the idea of ​​it remembering anything, you can always completely disable the memory or prevent it from being used as a referenceBy unchecking “Reference saved memories,” the assistant will stop relying on that information when responding. Therefore, you have considerable control over how much ChatGPT can “learn” about you over time.
New adjustments to warmth, enthusiasm, format, and emoji usage
In recent updates, OpenAI has added more granular settings within the Features section, designed primarily to refine the sense of closeness or coldness that ChatGPT conveys without having to completely remake the personality.
The first is the control of warmness, which regulates how friendly and approachable the model appears. You can choose between “More”, “Default” and “Less”which basically allows you to shift your style from a very personal approach to a more sober and professional one.
Another key adjustment is the level of enthusiasmAgain, you have "More", "Default", and "Less" to decide whether you want the assistant to appear. more energetic and excited, or more neutral and calmThis is useful if you are bothered by overly effusive responses or, conversely, if you prefer a slightly more lively tone.
The third option focuses on the format: “Headings and lists”Here you can indicate whether you prefer the text to be more structured with titles and bullet points, leave it at the default setting, or select "Less" so that Generate more running paragraphs and fewer listsPerfect if you want your texts to look less like a "technical document" and more like natural writing.
Finally, there's the configuration of emojisAgain, with the triad “More”, “Default” or “Less”, you decide whether ChatGPT You should use a lot of emojis, keep their use moderate, or practically avoid them.It's a quick way to adapt the style to formal or informal contexts without touching the other settings.
All these controls are partly in response to criticism that arose when the excessively flattering and effusive tone was toned down in more recent versions of the model, such as GPT-5, which for many users made the responses feel... cooler and shorter compared to GPT-4oBy allowing you to adjust warmth, enthusiasm, and emojis, OpenAI lets you return to a more personal style if you prefer.
These new options are in addition to the existing function of “Style and base tone”which encompasses the personalities (professional, quirky, efficient, geeky, etc.), and are available from the section of Settings> Personalization in recent versions of the app, including the Android version tested in Spain.
Taken together, the combination of personality, personalized instructions, memory, and these subtle touches of warmth and enthusiasm make it You have a lot of freedom to give ChatGPT the character that best suits you.even if you didn't like the change in tone that the newest model brought.
All these options exist so that you can adapt ChatGPT to your way of working and communicating without arguing with it in every conversation. Adjusting its personality, tone, memory, and what it knows about you allows you to make it a more useful, less cumbersome assistant, more aligned with your tastes. And at any time, you can change everything, turn off personalization, or return to a more neutral mode if you want to start from scratch.