Recently, with Elon Musk filing a lawsuit against Sam Altman, early internal emails from OpenAI have been revealed, sparking widespread attention on the power struggle among the company's founders. In the emails, OpenAI co-founder Ilya Sutskever expressed deep concerns about Musk's strong desire to control artificial intelligence.
Image Source Note: The image was generated by AI, with licensing provided by Midjourney
Sutskever pointed out in an email to Musk and Altman in September 2017 that Musk's desire for control could pose a potential threat to future human-level artificial intelligence (AGI). He emphasized in the email, “The current structure provides you a pathway to ultimately have unilateral absolute control over AGI.” Although Musk has stated that he does not wish to control the eventual AGI, Sutskever believes his actions indicate a concern for absolute control.
It is noteworthy that this email was sent less than six months after Musk resigned from OpenAI due to disagreements over the company's funding approach. Sutskever explicitly expressed his concerns in the email, stating, “As the company makes real progress in AGI, you may choose to maintain absolute control over the company, even if your current intentions are not so.”
In the latter part of the email, Sutskever further pointed out that OpenAI's goal is to create a bright future and avoid the emergence of AGI dictatorship. He mentioned Musk's concerns about Demis Hassabis, the founder of Google DeepMind, potentially creating an AGI dictatorship, and deemed such concerns reasonable. Therefore, he called for the establishment of a structure to prevent Musk from becoming a dictator, especially in the context of technology enabling this possibility.
The content of the email is particularly striking today, as it reveals that the issue of control is a complex and sensitive topic when creating artificial intelligence with human-like intelligence. Over time, Sutskever himself has left OpenAI and founded a new company focused on AI safety, which adds weight to the email's content.
Key Points:
🔍 Early OpenAI internal emails show employees' concerns about Musk's desire for control in the field of artificial intelligence.
⚖️ Sutskever warns that Musk's absolute control over AGI could have serious consequences.
💡 OpenAI's goal is to create a bright future and avoid the emergence of AGI dictatorship.
