Internal OpenAI Emails Show Employees Feared Elon Musk Would Control AGI

“The goal of OpenAI is to make the future good and to avoid an AGI dictatorship.”

Absolute Power

During the discovery process in Elon Musk’s lawsuit against Sam Altman, email exchanges from early in the group’s history show that even early on, tensions flared over who would control the company’s powerful creations.

In one of these early emails submitted as evidence exhibits in the Musk vs. Altman trial, OpenAI cofounder Ilya Sutskever took him to task for his egoistic need for control — and the dangers it could pose for any forthcoming human-level AI, better known as artificial general intelligence (AGI).

“The current structure provides you with a path where you end up with unilateral absolute control over the AGI,” Sutskever wrote to Altman and Musk in September of 2017. “You stated that you don’t want to control the final AGI, but during this negotiation, you’ve shown to us that absolute control is extremely important to you.”

“As an example, you said that you needed to be CEO of the new company so that everyone will know that you are the one who is in charge,” he continued, “even though you also stated that you hate being CEO and would much rather not be CEO.”

Under Control

Notably, the email was sent less than six months before Musk resigned from OpenAI over disagreements about how the company should raise money — which is also the crux of his lawsuit against Altman et al now.

“We are concerned that as the company makes genuine progress towards AGI, you will choose to retain your absolute control of the company despite current intent to the contrary,” Sutskever wrote.

Similar concerns may well have inspired Sutskever to lead a briefly successful coup against Altman last year before his own apparent ouster this spring. As the rest of his scathing email to Musk shows, he had good reason for worry.

“The goal of OpenAI is to make the future good and to avoid an AGI dictatorship,” he wrote. “You are concerned that Demis [Hassabis, the founder of Google’s DeepMind AI lab] could create an AGI dictatorship. So [are] we. So it is a bad idea to create a structure where you could become a dictator if you chose to, especially given that we can create some other structure that avoids this possibility.”

Reading the message in hindsight — especially after Sutskever’s own exit and founding of a new venture promoting AGI safety — is pretty chilling, especially as Musk’s embrace of embrace of president-elect Donald Trump reveals a deep thirst to control how the world is run.

More on Musk’s control issues: Elon Musk’s Daughter Vivian Says He’s a “Delusional and Grubby Little Control Freak”