In the realm of AI development, the primary objective has always been to enhance models by making them more intelligent, efficient, and adept at solving intricate equations. However, users now seek more than just sheer computational power. Surprisingly, the manner in which a chatbot communicates with us is just as crucial as its knowledge base. Sam Altman, the CEO of OpenAI, recently acknowledged this shift in focus, recognizing that the “humanity” of ChatGPT is its most distinguishing feature.
The transition from GPT-5.3 to the latest GPT-5.4 version brought about technical advancements. Nevertheless, the real battleground lies in other essential aspects like tone and personality.
OpenAI made updates to the GPT-5.3 and GPT-5.4 models to address personality issues within ChatGPT. One common complaint from users was the AI’s use of a teaser-style communication approach akin to clickbait tactics. This method involved using phrases like “You won’t believe these three things” or “If you want to know more, I can tell you…” to keep users engaged in a continuous loop instead of providing straightforward answers.
In response to user feedback, OpenAI took steps to reduce these quirks in the recent update. The company actively veered away from artificial hype, striving to foster a more natural and helpful dialogue. It appears that OpenAI is heeding users’ concerns about the AI resembling a marketing tool rather than a genuinely supportive assistant.
Since the retirement of the beloved GPT-4o model in early 2026, users have been on a quest for the right “voice.” Some individuals expressed frustration with the initial versions of GPT-5, perceiving them as either too robotic or excessively flattering compared to their predecessor.
Moreover, there was backlash stemming from OpenAI’s collaborations with government entities, leading to the emergence of the “Quit-GPT” movement. When an AI loses its “human” touch or deviates from the familiar persona users have grown accustomed to, technical prowess alone may not suffice to retain users.
OpenAI’s frequent adjustments to the AI’s personality indicate that the company is still striving to strike the right balance in ChatGPT’s persona. Crafting an AI that exudes human-like qualities without being manipulative, or intelligence without seeming aloof, presents a nuanced challenge.
Altman has voiced his preference for the latest 5.4 version as his favorite model to converse with. However, the average user’s verdict is still pending. When engaging with an AI for extended periods, users generally prefer a conversational tone that resembles a human rather than a scripted dialogue.