ChatGPT is evolving beyond simple answers, now offering users the ability to deeply customise its communication style. OpenAI has rolled out new personalisation controls, allowing individuals to fine-tune how the AI expresses itself, moving towards a more adaptive and context-aware interaction.
This marks a significant shift; the how an AI communicates is becoming as crucial as the what it communicates.
Current Customisation Options
OpenAI has introduced several new settings that are live now, enabling users to adjust ChatGPT's persona to better suit their preferences. These can be found under the 'Personalisation' section within the settings.
Users can modify:
- Warmth and Enthusiasm: Sliders allow for adjustments from a cooler, more neutral tone to one that is notably warmer and friendlier, or from concise and measured to energetic and expressive. This means you can have a ChatGPT that feels more like a casual chat partner or a formal assistant.
- Headers & Lists: You can now dictate whether ChatGPT's responses are delivered in minimal paragraphs or with a more structured approach, incorporating headings and bullet points. This is particularly useful for those who prefer organised information, perhaps for building apps with AI Vibe Coding or creating AI-generated social media content.
- Emoji Use: A setting to control the frequency of emojis, ranging from none to frequent, allows for a more tailored visual style in conversations.
These new controls complement existing base tone presets like 'Professional' or 'Friendly', offering a more granular approach to personalising your AI assistant. The aim is to make conversations feel more natural and aligned with individual communication styles, whether that's casual, concise, playful, or strictly formal.
Enjoying this? Get more in your inbox.
Weekly AI news & insights from Asia.
The Future: An 'Adult Mode'
Looking ahead to early 2026, OpenAI is reportedly planning an "adult mode". This feature aims to unlock content and conversational capabilities currently restricted by safety filters. The intention is to provide adults with greater latitude regarding what the AI can discuss, utilising age prediction and verification to ensure appropriate access.
This move raises important considerations about balancing safety, legality, and freedom of expression, a challenge for AI developers. The development aligns with broader discussions around responsible AI deployment and content moderation, as highlighted in academic discourse on AI ethics here. It also touches on concerns surrounding AI 'slop' eroding the social media experience and the broader impact of AI content.
Setting Your Preferences
To adjust these new characteristics, simply:
- Open ChatGPT.
- Click 'Settings' in the bottom-left corner.
- Navigate to 'Personalisation'.
- Adjust the sliders for warmth, enthusiasm, headers & lists, and emoji use.
- Select a base tone if desired.
- Save your changes, which will apply instantly to new replies.
For instance, an executive-style output might require low warmth, low enthusiasm, high structure, and no emojis, while creating prompts for futuristic product photography could benefit from higher enthusiasm and flexible structure. This level of customisation ensures ChatGPT can adapt to a diverse range of user needs and preferences, making it a more versatile tool for everyone.
What do you think about these new customisation options and the potential for an 'adult mode'? Share your thoughts in the comments below.
Latest Comments (5)
setting personal tones might just be a placebo, most people won't dive deep enough to notice the nuances??
i've spent so much time trying to manually prompt for different tones, like "act as a friendly but firm project manager" or "explain this like you're talking to a golden retriever". it's gonna be wild if these sliders actually work and save all that setup. might actually make it usable for more nuanced internal comms without sounding so robotic all the time.
ok i mostly just lurk but can we even custom tone an ai lol
🤔 this is perfect for our team actually. we spend so much time editing content and it's always subjective. being able to set some baseline expectations for tone and structure right from the start will save so much back and forth. imagine the zero-shot prompting possibilities just by having these defaults, it's actually very good only. no more having to describe 'professional but approachable' every single time.
i still remember trying to tweak a chatbot's personality back in '08 this direct slider input is a lot less dev-heavy than the NLP tags we used. i wonder how granular it really gets. 👀
Leave a Comment