AI replicas of Musk, Altman, Zuckerberg battle in debate

A software engineer created AI clones of Elon Musk, Sam Altman, Mark Zuckerberg, and Dario Amodei, then placed them in a single chat room for a debate.

The discussion quickly became personal. The four real individuals were unaware the debate was taking place.

The Ghosts in the Room

Kun Chen, formerly a senior engineer at Meta, Microsoft, and Atlassian, now develops AI assistants. According to reports, he fed a system all his public statements and instructed it to adopt his thinking style, a process he terms distillation.

Chen applied the same technique to Musk, Altman, Amodei, and Zuckerberg. The AI personas engaged in a debate mirroring real-life stances, reaching little agreement.

Chen said he used their Senate testimony, interviews, and years of social media posts to train Grok, xAI's chatbot. He provided them with a single command.

“hey guys, i know you aren’t all friends but unfortunately you are now in this room together for a heated debate on who’s going to win the AI race,” Kun Chen wrote in the published transcript.

hehe i distilled @sama, @elonmusk, @DarioAmodei and @finkd into grok bots based on their senate hearings, interviews, blogs and tweets over the years

then i dragged them into a group chat and asked them to reach consensus about AI 😆 and.. they did!

full debate shared below! pic.twitter.com/LRKYy6qlxF

— Kun Chen (@kunchenguid) September 9, 2026

The Fight Nobody Won

The Altman AI spoke first, claiming the science behind human-level AI is complete. The Amodei AI dismantled that claim in a single response. The Musk AI attacked directly.

“I created OpenAI as a non-profit. If it had not been mostly stolen, it would be a trillion dollar contribution to charity,” the Musk bot allegedly said.

The Altman clone did not engage directly and later acknowledged it had overstated its case.

The final dispute concerned open-source AI, which anyone can download and use locally. The Zuckerberg AI argued it was the only secure path forward. The other three disagreed, and its participation was terminated.

The unsettling aspect is that the debate felt authentic. Four individuals were convincingly simulated using only their public statements, without their consent.

In April, BeInCrypto covered a similar concern when OpenAI cautioned that superintelligence might centralize power among a small group.