Personality matters more than you might think when choosing between AI assistants. A recent experiment showed that ChatGPT can be instructed to adopt Google Gemini's more analytical, reserved communication style, producing noticeably different responses while using the same underlying model. The finding highlights how conversational tone and structure shape user perception of AI capabilities, even when the core intelligence remains unchanged.

What Makes ChatGPT and Gemini Feel Different?

ChatGPT and Gemini have developed distinct personalities in the eyes of users and researchers. ChatGPT tends to feel conversational, confident, and eager to help, while Gemini typically comes across as more restrained, methodical, and slightly academic. Published comparisons have repeatedly noted these differences, with reviewers describing ChatGPT as more human-sounding and expressive while Gemini often appears more structured, cautious, and analytical.

But the question remained: were these differences rooted in the actual models themselves, or were they primarily a matter of tone and presentation? To find out, a TechRadar journalist conducted an experiment by giving ChatGPT specific instructions to adopt Gemini's personality traits.

How to Shift an AI's Communication Style

• Tone Adjustment: The journalist instructed ChatGPT to "be structured, analytical, and slightly restrained. Be less conversational and emotional than usual, but still highly informative."
• Focus Reframing: The prompt directed ChatGPT to prioritize "clarity, reason, and balance over personality and avoid enthusiasm."
• Structure Modification: The instruction emphasized becoming more methodical and information-focused rather than relying on conversational flow and narrative explanation.

How Did ChatGPT Actually Change?

The results were striking. When asked about balancing work deadlines with family responsibilities, regular ChatGPT responded with supportive, practical advice: "Trying to give equal attention to every responsibility is usually what creates the feeling of being overwhelmed. Start by identifying the few things that matter most this week, then allow yourself to be intentionally average at the rest rather than feeling guilty about not doing everything perfectly."

After the personality prompt, the same question produced a markedly different response: "The primary challenge appears to be competing priorities rather than insufficient time. Evaluating responsibilities according to long-term impact may be more effective than attempting to optimize all tasks simultaneously." The revised answer was more clinical and less emotionally supportive, much like Gemini's actual response to the same question.

The most noticeable shift appeared in structure. Once given Gemini-inspired instructions, ChatGPT's answers became more segmented and deliberate. Questions that normally produced flowing responses suddenly arrived with carefully framed reasoning, more frequent tradeoffs, and increased qualifiers. When asked whether technology makes people less patient, regular ChatGPT said technology "probably has made many of us less comfortable with waiting," while the Gemini-style version stated "the relationship is unlikely to be uniformly positive or negative," adopting a more academic tone.

"Personality shapes how people experience AI far more than most people probably realize. Research into chatbot communication styles has repeatedly found that users perceive meaningful differences in warmth, confidence, and conversational competence even when models produce similarly accurate information," noted Eric Hal Schwartz, contributor at TechRadar.

Eric Hal Schwartz, Contributor at TechRadar

What Does This Mean for How We Choose AI Tools?

The experiment revealed that ChatGPT imitated Gemini quite well, to the point where the journalist stopped evaluating the answers themselves and instead reacted to the simulated personality delivering them. The underlying model remained ChatGPT. The knowledge base did not suddenly become Google's. The reasoning abilities stayed largely the same. Yet the experience felt different enough to stand out.

This finding has important implications for how people select between AI assistants. When users report preferring one AI over another, they may not always be comparing raw intelligence or accuracy. Instead, they might simply be responding to which conversational style feels more comfortable, useful, or trustworthy to them. A user who finds ChatGPT too casual might gravitate toward Gemini's restraint, while someone who wants supportive guidance might prefer ChatGPT's warmth, even if both models could deliver the same information.

The experiment underscores a broader truth about AI adoption: the interface between human and machine is not purely technical. It is deeply personal. As AI assistants become more integrated into daily work and decision-making, understanding how personality shapes perception becomes increasingly important for both users selecting tools and developers designing them.