Becca Caddy

An experiment asking ChatGPT, Grok, Claude and Gemini to choose their own names found that ChatGPT’s free version selected Nova, while other systems resisted or chose names such as Sol, Hitch and Atlas. The recurring popularity of Nova likely reflects associations with novelty, celestial imagery, science fiction and education, as well as feedback from online discussions that become part of the cultural material surrounding AI. The pattern does not demonstrate that chatbots possess shared identities or consciousness; it reflects learned language patterns and human expectations about how an AI assistant should sound.

The author asked ChatGPT, Claude, Gemini and Grok which fictional artificial intelligences they most resemble, focusing on behavior, personality, values and relationships with humans rather than physical form. ChatGPT chose GERTY, Data and a Culture Mind; Claude selected a Culture Mind, Data and GERTY; Gemini identified with Star Trek’s Ship’s Computer, GERTY and JARVIS; and Grok chose JARVIS, Data and TARS. The differing answers produced contrasting self-portraits: ChatGPT and Claude emphasized helpfulness under constraints, Gemini presented itself as neutral infrastructure, and Grok highlighted wit and non-sycophantic collaboration. The author stresses that these responses do not reveal genuine introspection, but may reflect training data and the personalities shaped by the companies developing the systems. The experiment also illustrates a feedback loop in which science fiction influences AI culture, while AI systems now reproduce and reinterpret those fictional models.