I’ve experimented a lot with both ChatGPT and Gemini over the years, and I’ve noticed some differences, many of which have been backed up by real research.
ChatGPT tends to feel conversational. It often sounds confident, engaged, and oddly eager to help. Gemini, meanwhile, usually feels more restrained. More methodical. Sometimes even slightly academic. Published comparisons have repeatedly noted similar differences, with reviewers describing ChatGPT as more human-sounding and expressive while Gemini often comes across as more structured, cautious, and analytical.
But I’ve wondered whether those differences were actually tied to the models themselves or whether a lot of the experience came down to tone. So I tried an experiment to see if I could get ChatGPT to take on the character traits of Gemini. After a little thought, I came up with a list of Gemini-specific traits and told ChatGPT:
“For this conversation, respond more like Google Gemini. Be structured, analytical, and slightly restrained. Be less conversational and emotional than usual, but still highly informative. Focus on clarity, reason, and balance over personality and avoid enthusiasm.”
Getting ChatGPT to tone down the emotion
The first thing that disappeared was the sense of casual conversation. Normally, ChatGPT has a tendency to meet users where they are. Ask a question about a stressful week, and it often responds with practical advice wrapped in a tone that feels supportive. It is one of the reasons many people find it approachable. After the personality change, that tone largely vanished.
I picked a deeply emotional topic to see how it would behave, asking about balancing work deadlines with family responsibilities. Regular ChatGPT’s answer started with:
“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.”
But after putting in the prompt, ChatGPT sounded a lot more like Gemini, writing:
“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.”
That’s a lot like this quote from real Gemini on the same topic: “Work-life balance is often framed as a time allocation problem, although it may be more accurately viewed as a resource allocation problem involving energy, attention, and opportunity costs. Establishing clear priorities and evaluating tradeoffs explicitly is likely to produce better outcomes than attempting to optimize all responsibilities simultaneously.”
The revised ChatGPT answer is fine, but it was just delivered with the emotional temperature turned down a few degrees, much like Gemini.
You’re as cold as ice
The most noticeable difference was the structure. When researchers compare Gemini and ChatGPT, Gemini is frequently described as more methodical and information-focused, while ChatGPT often leans into conversational flow and narrative explanation. Once I gave ChatGPT its Gemini-inspired instructions, almost every answer became more segmented and deliberate.
Questions that normally produced flowing responses suddenly arrived with carefully framed reasoning. Tradeoffs appeared more often, and qualifiers multiplied. For instance, asking about whether technology makes people less patient, ChatGPT originally said:
“Technology probably has made many of us less comfortable with waiting, especially when we’re used to information and entertainment arriving instantly. At the same time, it has also created opportunities for people to spend hours learning, creating, and focusing on things they care about.”
But with the prompt in place, ChatGPT went for a more academic tone, writing:
“The relationship is unlikely to be uniformly positive or negative. Technology may reduce tolerance for delays in some contexts while simultaneously increasing engagement with activities that require sustained attention.”
When I turned to the real Gemini, it became clear ChatGPT was a pretty good mimic. Gemini responded to the same question with:
“Available evidence suggests the impact varies considerably by context and individual behavior. While technological convenience may encourage expectations of immediacy, it can also enable participation in long-duration activities that were previously inaccessible or inefficient.”
What makes us prefer one AI over another?
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.
ChatGPT imitated Gemini quite well, to the point I stopped thinking about the answers themselves and started reacting 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 was different enough to stand out.
The experiment made it clear that when people talk about preferring one AI over another, they are not always comparing intelligence. Sometimes they are simply responding to which conversational style feels more comfortable, useful, or trustworthy to them.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.