Sam Altman claims AI will run research labs by 2028

OpenAI CEO Sam Altman has unveiled the company’s most ambitious goal yet, i.e., to build a fully autonomous “AI researcher” by 2028.

Speaking during a recent livestream, Altman said OpenAI plans to first develop an “intern-level research assistant” by 2026, before advancing to a system capable of conducting complex scientific research without human oversight two years later.

The announcement comes as the company officially transitions into a public benefit corporation (PBC), a move that allows it to raise capital more freely while maintaining its social mission.

Altman explained that the new structure gives OpenAI the flexibility to fund massive computational projects that were previously out of reach under its non-profit model.

Moreover, OpenAI’s chief scientist, Jakub Pachocki, elaborated that the upcoming AI researcher will not only mimic human behaviour but will be able to independently execute research projects from start to finish.

He added that the company’s long-term vision includes achieving “superintelligence,” potentially within the next decade.

It is worth mentioning that to make this vision a reality, OpenAI is investing heavily in two areas: algorithmic innovation and increased “test-time compute,” which allows AI models to dedicate vast computing resources to solving a single scientific challenge.

Altman also revealed that OpenAI is building 30 gigawatts of computing infrastructure, valued at $1.4 trillion, to support its next generation of AI models.