OpenAI will have 'legitimate AI researcher' by 2028, Sam Altman

OpenAI CEO Sam Altman said during a livestream on Tuesday that their AI deep learning system are rapidly improving with capabilities to solve complex tasks faster. In fact, so fast that OpenAI is tracking toward achieving an intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028.

OpenAI has completed its conversion to a public benefit corporation, shifting away from its original non-profit framework. This move coincides with an ambitious timeline for the organization. The restructuring allows OpenAI to overcome previous limitations associated with its non-profit charter and unlocks new avenues for raising capital.

During the livestream, Jakub Pachocki, OpenAI’s chief scientist, joined Altman and described that this AI researcher should not be compared with a Human who researches AI, as a “system capable of autonomously delivering on larger research projects.”

“We believe that deep learning systems may be less than a decade away from superintelligence,” Pachocki added. He portrayed superintelligence as systems more intelligent than humans across a large number of critical actions.

OpenAI aims to achieve its goals through two primary strategies: ongoing algorithmic innovation and a significant increase in “test time compute,” which refers to the duration models spend contemplating problems. According to Pachocki, current models can manage tasks with approximately a five-hour time horizon and compete with top human performers in events like the International Mathematical Olympiad.

He anticipates this horizon will expand rapidly, partly by enabling models to allocate substantially more computational resources to complex problem-solving. For significant scientific advancements, he suggested that it would be worthwhile to dedicate the computing power of entire data centers to a single problem.

According to OpenAI, these objectives are in queue with the firm’s overall push to advance scientific research and allow AI to potentially make discoveries faster than human researchers, solve problems beyond current human capacities, and drastically hasten technological innovation across numerous fields like medicine, physics, and technology development.

Altman also mentioned that reorganizing creates a framework to support OpenAI’s bold timeline for AI research assistants while maintaining a commitment to reliable AI development. Under this new structure, the non-profit OpenAI Foundation, which is concentrated on scientific advancement, will own 26% of the for-profit and will oversee the research direction.

Altman states that the for-profit division’s capacity to secure additional funding is crucial for scaling the infrastructure required to achieve scientific breakthroughs.

“OpenAI has devoted to 30 gigawatts of infrastructure, which is a $1.4 trillion financial obligation, over the next few years,” Sam Altman added.