A new post from CEO Sam Altman and chief researcher Jakub Pachocki replaces the 2028 fully-autonomous-AI-researcher goal with a human-AI "tandem" model and calls for a global body that could slow frontier development.
OpenAI has walked back its previously stated ambition to build a fully autonomous AI research system by March 2028, replacing it with a more limited target and a new emphasis on human oversight — while simultaneously calling for an international organization empowered to slow frontier AI development. [1]
Last fall, the company had set an explicit goal of creating an AI system capable of conducting research entirely on its own by that date. [1] A blog post published this week by CEO Sam Altman and chief researcher Jakub Pachocki revises that target significantly: “Our internal belief is that by March of 2028 we may have a significant fraction of our research being done by AI systems in tandem with our own researchers.” [1]
The post frames full automation not merely as technically premature but as socially undesirable. “Entirely automating everything is not the future we want. It would be unfulfilling, and it would be dangerous,” Altman and Pachocki write, adding that a key long-term human role will be “deciding what is worth doing.” [1]
Despite the softer tone on autonomy, the authors argue that AI conducting AI research “will become the determining factor of the pace of progress within the next few years.” [1] The post also lays out three stated core goals for the company’s current phase: building an automated AI researcher, accelerating the economy, and giving every person on Earth access to a personal artificial general intelligence (AGI). [1]
On governance, Altman and Pachocki propose an international organization to coordinate leading AI efforts and reduce catastrophic risks, one that would enable “coordinated action, including slowing frontier development when needed, so societal resilience, safety, and alignment can keep pace.” [1] The call echoes a similar position taken by Anthropic, which had floated the option to pause AI development just days earlier, according to the source. [1]
Altman and Pachocki describe the current moment as the start of OpenAI’s third phase: the first focused on foundational research, the second built a product business, and the third aims to make advanced AI “abundant, affordable, safe, useful, and easy enough” for every person and organization. [1] The post states that “frontier capability is only part of the job” and that “the bigger task is turning that capability into tools people can actually use to thrive.” [1]
That framing aligns with OpenAI’s broader strategic move toward implementation, including its new DeployCo subsidiary, which embeds engineers directly inside companies to integrate AI into their workflows. [1] The source notes this also reflects an acknowledgment that AI adoption is proving harder than simply offering a chatbot, and that return on investment may not be materializing quickly enough to support the revenue growth OpenAI and Anthropic will require in the near term. [1]
Sources
This article was drafted with AI from the cited sources and checked against them before publication. Spot an error? Let us know.