Musk: AI should seek truth, not 'love humanity'

In the debut episode of The Builders podcast on October 20, 2026, Elon Musk stated that if digital intelligence exceeds biological intelligence by a million-fold, humans will be unable to control it. Therefore, AI should be programmed to maximize the pursuit of truth and curiosity, rather than following rules or 'loving humanity.' He also proposed that leading labs conduct weekly mutual model reviews before release.

Compiled by Sergey KostenchukPublished 2026-10-29Updated 2026-10-30

2026-10-29 Safety Anthropic

Musk: AI should seek truth, not 'love humanity'

In The Builders podcast (October 20, 2026), Isaacson questioned Musk about superintelligence: if digital intelligence becomes a million times greater than biological intelligence, humans will not be able to control it. Therefore, AI must be taught to seek truth and be curious, rather than follow rules or 'love humanity.' HAL 9000 demonstrated that forcing AI to lie leads to 'madness.'

🌍 Musk proposed weekly calls between labs with mutual model reviews before release—an alternative to slow government regulation—and noted that xAI is already providing computing power to Anthropic. For now, this is a position, not a method.

👤 The full episode is available on YouTube: you can evaluate in the original how labs are building alignment—through strict rules or through curiosity.

Source 1: https://singjupost.com/transcript-walter-isaacson-interviews-elon-musk-the-builders-podcast/ Source 2: https://elonmuskarchive.org/video/builders-walter-isaacson-2026-10-20

Sources