Challenges in Personalising AI: From Uniformity to Uniqueness

Large language models (LLMs) currently adopt a one-size-fits-all approach, neglecting the rich diversity of individual personalities and communication styles. This article investigates both the promise and the challenges of developing personality-aware AI systems. It critically examines how established alignment methods — such as Reinforcement Learning from Human Feedback (RLHF) — favour broad acceptability over personalisation, and explores strategies like prompt-based adaptation and user-specific tuning. Addressing technical hurdles including the cold start problem and long-term memory constraints, the discussion also delves into ethical considerations surrounding privacy, fairness, and bias. Ultimately, the article argues that achieving meaningful personalisation in AI will require a balanced framework that enhances user engagement and trust while upholding societal values and transparency.