'Lacked technical nuance': OpenAI CEO responds to AI safety concerns raised by Elon Musk and others

OpenAI CEO Sam Altman has responded to the open letter signed by Elon Musk and hundreds of others in the tech industry concerning concerns over AI. Altman noted at an MIT event that the letter “lacked technical nuance about where we need the pause” and called for more detailed consideration of AI safety.

Released last month, the letter called for a six-month pause on the development of AI models that are “more advanced” than OpenAI's GPT-4, the latest version of the popular chatbot ChatGPT. Altman agreed that the safety of AI models should be carefully considered before release, noting that OpenAI spent more than six months studying GPT-4 before its release.

"We spent more than six months after we finished training GPT-4 before we released it, so taking the time to really study the safety of the model,” Altman said at MIT.

Altman also stated that OpenAI is not currently training GPT-5, contrary to the earlier version of the letter and that the company will continue to be as transparent as possible about its AI developments. As the capabilities of AI become more advanced, Altman emphasized that the safety bar must also increase.

The letter has been criticized by some, including tech giants like Bill Gates and Google CEO Sundar Pichai, who claim that a pause is impractical and would require government involvement to enforce. Others, such as LinkedIn co-founder Reid Hoffman, claim that Musk's signature on the letter is motivated by jealousy.

Altman stressed that AI development impacts everyone and that it is essential to engage as many people as possible in testing and learning about them. Despite the challenges and frequent need to change direction with new data, he believes it is worth the tradeoff to put imperfect AI systems out into the world for public discussion and experience.

Also Read

Instagram revamps Reels with new video editing and discovery features

Europe sets up task force on ChatGPT to create a common policy on AI privacy rules

Meta’s new AI project turns doodles into animated figures