OpenAI has transformed ChatGPT from a one-size-fits-all assistant into a customisable conversational partner. The new personalisation controls, launched in December 2025, allow users to fine-tune everything from warmth levels to structural formatting. This shift represents more than technical enhancement. It signals AI's evolution towards truly adaptive communication that matches user preferences and working contexts. For Asian users navigating AI interactions across multiple languages, cultural expectations, and professional contexts, the tone customisation controls provide meaningful flexibility that has been missing from most AI assistants.
The implications extend far beyond casual chatting. As AI becomes integral to professional workflows, from prompt engineering to business communication strategies, the ability to match communication style to context becomes increasingly important. An AI assistant that is appropriately warm for customer service but appropriately formal for legal drafting serves both use cases better than a single default tone applied to all interactions.
Fine-tuning your AI companion
The new controls live within ChatGPT's personalisation settings and offer unprecedented granularity. Users can adjust warmth from cool and neutral to notably friendly, while enthusiasm ranges from measured and concise to energetic and expressive. Structural preferences matter equally. The headers and lists toggle determines whether responses arrive as flowing paragraphs or organised bullet points.
For professionals creating structured content workflows, these controls reduce the need for repetitive prompt engineering. Setting preferred structure once means subsequent queries follow that structure automatically. Setting preferred tone once eliminates the need to specify tone in every query. The time savings accumulate across thousands of interactions.
The controls interact with ChatGPT's memory system in useful ways. ChatGPT learns preferences over time and applies them contextually. A user who typically requests formal responses but occasionally asks for casual creative writing can have both needs accommodated as ChatGPT learns which context calls for which style. OpenAI's product blog has documented the personalisation controls and their interaction with memory.
Why tone customisation matters for Asian users
Asian users often have specific tone preferences that default AI tone settings fail to match. Japanese users frequently expect more formal register in professional interactions, with specific honorific conventions that English-trained AI often handles awkwardly. Korean users expect similar honorific appropriateness. Chinese users have varying preferences depending on context, with business communications typically more formal than casual interactions.
Indian English users have specific regional conventions that differ from both American and British English defaults in AI systems. Indian English tends to be more formal than US English and uses specific phrasing that native Indian English speakers prefer. Tone customisation allows users to adjust responses toward their preferred register.
Southeast Asian users span diverse language preferences. Users writing in English alongside Bahasa Malaysia, Bahasa Indonesia, Tagalog, or Thai expect responses that respect both language formalities. The tone customisation interacts with language preferences to produce responses appropriate for specific cultural contexts.
The technical implementation behind the controls
The personalisation controls work through a combination of system prompt adjustments and ongoing training data that teaches the model how to interpret different tone settings. When users adjust warmth or enthusiasm, the underlying API request to ChatGPT includes system-level instructions that calibrate response generation accordingly.
The approach is different from simple prompt prefix injection. ChatGPT's personalisation controls are integrated into the model's training such that different settings produce meaningfully different response styles rather than cosmetic differences in word choice. OpenAI has indicated that specific fine-tuning work went into making the controls produce genuinely different response characteristics rather than surface-level variation.
User preferences are stored securely and applied consistently across sessions. Users can reset preferences, adjust them as needed, or maintain different preferences for different use cases through ChatGPT's workspace features. The system design balances persistence (remembering preferences) with flexibility (allowing adjustment when context changes).
Professional workflow implications
For professional users, the tone controls enable specific workflow optimisations. Business writing benefits from formal, structured settings. Creative writing may benefit from warmer, more enthusiastic settings. Technical documentation calls for structured formatting with headers and lists. Legal drafting benefits from measured, precise tone without unnecessary warmth.
Teams collaborating on shared projects can establish consistent tone preferences across team members' ChatGPT use, ensuring uniformity of AI-assisted outputs. Enterprise deployments through ChatGPT Business and Enterprise tiers include organisational defaults that can be adjusted by individual users within team-level guardrails.
Specific use cases that benefit from customisation include customer service responses (warm, helpful, concise), technical support (clear, structured, actionable), internal communications (professional, appropriate formality), creative brainstorming (enthusiastic, exploratory), and research synthesis (measured, thorough, well-organised). Each calls for different tone and structure, and the customisation controls enable all these styles within a single AI assistant.
How this compares with competing AI products
Claude from Anthropic also offers personalisation features, though with somewhat different control mechanisms. Claude's constitutional AI approach produces responses with specific characteristic tone that users can adjust but within narrower bounds than ChatGPT's current controls. Some users prefer Claude's more consistent default behaviour; others prefer ChatGPT's greater adjustability.
Google Gemini offers personalisation through its integration with user Google profiles. Gemini's approach leverages data that Google already holds about user preferences to adjust AI behaviour. For users deeply integrated into Google's ecosystem, this contextual personalisation can produce appropriate responses without explicit control adjustment.
Chinese AI products including Doubao, ERNIE Bot, and Tongyi have developed personalisation controls tailored to Chinese user preferences. These include specific controls for formality level appropriate to Chinese business culture, length preferences consistent with Chinese content consumption patterns, and style preferences that match Chinese aesthetic conventions. Gartner research has documented how different AI products approach personalisation across markets.
The accessibility and inclusion dimension
Personalisation controls have significant implications for accessibility and inclusion. Users with specific cognitive preferences including those who prefer detailed structured responses versus those who prefer concise flowing prose can adjust AI behaviour to match their thinking style. Users who find AI default tones off-putting can adjust toward more comfortable interaction styles.
Non-native English speakers using ChatGPT may prefer simpler language, shorter sentences, or more literal phrasing. Tone controls can support these preferences alongside dedicated language support features. Users with specific learning differences may benefit from more structured, bulleted responses that support comprehension.
Elderly users, children, and users with limited technology experience may prefer warmer, more patient interaction styles. While ChatGPT's customisation does not fully solve inclusion challenges for all these user segments, it provides meaningful flexibility that benefits users who do not fit the default assumptions.
Enterprise and regulatory considerations
For enterprise deployments, personalisation controls raise specific considerations. Organisations want AI-generated content to reflect consistent brand voice and professional standards. Individual user customisation can conflict with organisational consistency requirements, particularly when AI-generated content is produced for external customers or partners.
ChatGPT Business and Enterprise tiers allow administrative controls that establish organisational defaults while permitting individual adjustment within defined bounds. This balance supports both organisational consistency and individual productivity. Specific industries including financial services, healthcare, and legal have additional considerations around consistency and compliance.
Regulatory considerations include ensuring that AI responses meet specific regulatory requirements regardless of tone setting. Financial advice regulations, healthcare communication standards, and legal practice rules all have specific requirements that must be met whether the AI response is warm or cool, structured or flowing. The customisation controls affect style but not substance, though users and organisations need to verify that outputs meet regulatory requirements.
What users should explore
For current ChatGPT users, experimenting with the personalisation controls can produce noticeable improvements in AI interaction quality. Users who have grown accustomed to the default tone may find that adjusted settings better match their preferences and working style. The OpenAI help centre provides guidance on specific controls and their effects.
For enterprise users, establishing team defaults for different use cases supports consistency while preserving individual flexibility. Pilot testing different tone and structure combinations for specific workflows identifies optimal settings before rolling out broadly.
For Asian users specifically, the tone customisation matters more than might be immediately apparent. Default AI tones trained primarily on American English conventions do not match all cultural contexts equally well. Explicit adjustment toward preferred tones produces responses that feel more appropriate and usable in non-American contexts.
The broader takeaway is that AI assistants are maturing into genuinely customisable tools rather than one-size-fits-all products. Personalisation is becoming a competitive differentiator across AI products. Users benefit from paying attention to customisation features because they genuinely affect how useful AI is for specific needs. OpenAI's December 2025 controls are one step in this maturation, and expect continued evolution as AI products compete on how well they adapt to individual users' preferences, contexts, and needs.