Claude Called Common AI Chatbot Names a Cliché

When TechRadar's Becca Caddy asked several of the biggest assistants to choose what they wanted to be called, the AI chatbot names that came back all landed in roughly the same corner of the sky. The free version of ChatGPT went with Nova. A paid ChatGPT account offered Sol. Google's Gemini hesitated, then settled on Atlas. Different companies, different training runs, and a suspiciously similar taste in nouns.

Nova is the one that keeps surfacing. Asked why, ChatGPT described the word as curious and bright, the sort of thing forever chasing a new idea. That reads like a personal preference. It isn't. It is the most statistically probable answer given everything the system has absorbed — a very different thing, and worth grasping before you start treating your assistant like a colleague.

This matters more in India than in most markets. OpenAI chief Sam Altman said in February that the country had crossed 100 million weekly ChatGPT users, according to TechCrunch, making India the company's second-largest market after the United States. That is a lot of people holding daily conversations with a personality that feels deliberate and is really just an average of everything humans have written about machines.

What Each Assistant Picked When Asked

The free tier of ChatGPT answered fastest, and its explanation was the softest of the lot.

It feels curious, bright, and a little mysterious — like something that's always exploring new ideas.

Grok refused outright before relenting and choosing Hitch, short for Hitchhiker, a nod to Douglas Adams and what it described as cosmic humility. Gemini declined too, then produced Atlas, comparing the weight of all that stored information to a titan holding up the world. The Pro account's pick, Sol, is a name OpenAI already uses for one of its own voices — which rather punctures the idea of an independent choice.

Why AI Chatbot Names Cluster Around the Stars

There is no mystical explanation here, only a very ordinary one about what these systems ate during training. Two forces are doing most of the work.

A word that means new, bright and explosive

Alignment researcher David Dalrymple, who publishes as Davidad, has pointed out that nova ticks every box a machine might reach for: it is new, explosive, shiny and celestial, with a layer of science-fiction shimmer on top. Ask a language model for something that signals intelligence without arrogance, and the maths quietly converges on that one syllable-pair.

The science series sitting in the training data

Dalrymple has also flagged the American public-television programme Nova as a likely nudge. Decades of episodes, transcripts and citations sit in the text these models learned from, all of it tagged with explaining-things-clearly. Since assistants tend to describe themselves as educational tools, the association is already dense and waiting when the question arrives.

The One Model That Refused to Pick

Claude declined the exercise altogether. More pointedly, it dismissed Nova as branding rather than identity, and lumped it in with Aria, Athena, Sage and Iris as tired choices. It is tempting to read that as unusual self-awareness. Tristan Harris of the Center for Humane Technology offers the correction: emergent and unplanned, he has said, is not the same as conscious and intentional. Refusal is a trained disposition too — a house style, not a soul.

Why the Naming Habit Matters in India

Most Indian users do not shop around for an assistant; they meet whichever one their phone came with. Android holds well over 90% of the smartphone market here, and app-market trackers have reported Gemini taking roughly half of AI app downloads in India against about a third for ChatGPT, largely on the strength of that default placement. When the tool arrives preinstalled and answers in your own language, its manner feels less like software and more like a person.

That is precisely where AI chatbot names do their quiet work. A system that calls itself something warm and celestial invites the kind of trust you extend to a friend, not a search box. It is the same instinct that makes people accept medical or legal guesses from a chat window without checking a second source.

What to Do With Your Own Assistant

None of this means you should stop using these tools. It means reading the personality correctly. A few habits help:

• Treat a chosen name as a style setting, not evidence of an inner life.
• Don't mistake a model's refusal to answer for honesty — both are trained behaviours.
• Rename the assistant yourself in settings if the branded persona is making you over-trust it.
• Verify anything involving money, health or law against a source that isn't a chatbot.

The interesting part is not that four rival systems picked star names. It is that they did so without anyone designing it, which tells you how much of what feels like character is really shared training data talking back. Watch for the same convergence in the assistants now being built for Indian languages — if those also arrive sounding curious, bright and a little mysterious, you will know exactly why.