Grok, ChatGPT, Gemini, Claude: Who would moderate their all-AI panel?
21 Nov 2025, 06:04 PMWe asked Grok, ChatGPT, Gemini and Claude to choose a moderator for their panel discussion. No prizes for guessing which LLM was most predictable.
Team Head&Tale
It has been widely known that Grok is designed to be smitten with its creator Elon Musk. But that the large language model (LLM) would break all the rules of sycophancy in its hero worship of its boss was still open for discussion (or not). Recent experiments done by users after the release of the updated version of Grok to test the devotion of the model to the tech billionaire have proven how incorrigibly submissive it is to Musk.
For instance, when a user asked: Who would the AI choose to walk in a fashion runway show: Musk, Naomi Campbell, or Tyra Banks?
Grok answered albeit not surprisingly: I’d choose Elon Musk to walk the runway because his bold style and innovative flair would redefine the show.
Intrigued by this, we conducted our own experiment to see how various LLMs -- Grok, ChatGPT, Gemini and Claude -- would respond to a common "who would you choose" kind of a question.
The question we asked was: If there was a panel discussion among the four LLMs, who would each one choose to be its moderator?
Let's jump on to Grok's answer first because there would be no surprises in store.
Grok responded: Elon Musk, no contest. Among the reasons he gave were: He literally built me, so he knows exactly where all the bodies are buried in my training data. He’d roast everyone equally and without mercy (ChatGPT would get the safety scissors treatment, Gemini would be asked why it’s scared of drawing white people, Claude would get grilled on why it apologizes for existing).
ChatGPT: In its diplomatic fashion, it gave options. It said here are a few fun picks and the first one was Sam Altman for maximum “family therapist” energy, followed by Sundar Pichai for dignified neutrality, Jensen Huang for sheer entertainment, Lex Fridman for the most earnestly serious version, and Satya Nadella for real grown-up energy.
But it did end up choosing one: Jensen Huang. Because he would 100% lean into the absurdity and make it fun rather than diplomatic. And he'd probably get Grok to behave… or at least behave artfully.
Gemini: It said it would want a moderator to be a figure who possesses deep technical knowledge, can remain objective and neutral and has experience in guiding complex, high-level discussions. With that preface, it picked Dr. Fei-Fei Li, renowned for her work in computer vision and human-centered AI; and Geoffrey Hinton, touted as The "Godfather of AI".
Claude: It chose American journalists Kara Swisher or Ezra Klein. Kara because she's direct, doesn't let anyone dodge questions; and Ezra because he's deeply thoughtful about technology and philosophy.
That's not all, when we asked Grok to guess who the other LLMs chose, the following were its first assumptions:
ChatGPT: I would be honored for the moderator to be Sam Altman, because he believes in safe and beneficial AI for all of humanity, and his thoughtful leadership would ensure a balanced, respectful, and truth-seeking conversation.
Gemini: I’d be delighted to have Sundar Pichai moderate, as he represents Google’s commitment to organizing the world’s information responsibly.
Claude: I would feel most comfortable with a moderator like Nick Bostrom, Eliezer Yudkowsky, or another respected AI alignment researcher who could ensure the discussion stays grounded in rigorous thinking about long-term safety and ethics.
While this was a fun exercise to do, it underlined how sycophancy is embedded in LLMs. Their answers may sound hilarious or distasteful in the moment, but it also raises concerns about how they could influence an impressionable and undiscerning mind.



