OpenAI co-founder Ilya Sutskever to start new AI company

One month after OpenAI’s co-founder and chief scientist Ilya Sutskever left his company, he released a new plan to create another AI startup. He was part of the company’s board of directors, which led CEO Sam Altman to step down last November.

On Wednesday, Sutskever unveiled his new venture, Safe Superintelligence Inc., with a bold statement. He declared, “We will pursue safe superintelligence in a direct, uncompromising manner, with a singular focus, a single goal, and a unique product. Our path will be paved with revolutionary breakthroughs, all thanks to our small, yet highly skilled team.” Sutskever has partnered with former OpenAI technical staff Daniel Levy and former Apple AI director Daniel Gross to bring this vision to life.

Sutskever did not detail what “one product” would look like. He only implied that his new team would develop a control system for safe artificial intelligence by saying to Bloomberg, “We mean safe like nuclear safety.” He also stated that his team would not get into anything else until a safe superintelligence system was invented.

Sutskever, born into a Jewish family in Russia, spent his childhood in Jerusalem, Israel, starting at age five. Afterward, he immigrated to Canada to study neuromorphic computers that mimic human brain networks under Emeritus Prof. Geoffrey Hinton at the University of Toronto, who has been well-known as the leader in AI deep learning. It has been said that he was greatly influenced by his professor, who believed that AI could harm humanity.


이기욱 기자 71wook@donga.com