OpenAI is no longer talking like a startup experimenting with chatbots. The company is now openly discussing a future where artificial intelligence helps automate AI research itself, while also positioning its technology as something that could reshape economies, governments, and everyday life.
In a newly published post titled “Built for broad benefit: our plan,” OpenAI CEO Sam Altman and research chief Jakub Pachocki laid out what they describe as the company’s “third phase.” The post mixes optimism, warnings, long-term AGI ambitions, and subtle acknowledgements that AI development is accelerating faster than many people may realize.
One statement in particular stands out.
OpenAI says it believes that by March 2028, “a significant fraction” of its own AI research could be performed by AI systems working alongside human researchers.
That is a remarkable thing for one of the world’s most powerful AI companies to say publicly.
According to OpenAI, AI systems helping conduct AI research is not just about moving faster. The company argues it may become necessary to solve difficult alignment and safety problems before artificial general intelligence becomes even more capable.
“We believe that AI doing AI research will become the determining factor of the pace of progress within the next few years,” the company wrote.
While OpenAI continues to present AGI as a tool that should benefit humanity, the post also reveals how aggressively the company is thinking about automation, scale, and global influence.
The company compares AI to electrification in the early 20th century, arguing that transformative technologies eventually reshape society in ways people initially struggle to imagine. OpenAI says AI could help people understand medical bills, start businesses, care for aging parents, make financial decisions, and accelerate scientific discoveries.
At the same time, the company insists it does not want a future where humans are fully replaced.
“Entirely automating everything is not the future we want,” OpenAI wrote. “It would be unfulfilling, and it would be dangerous.”
Still, some readers may find the messaging contradictory. OpenAI argues against concentrated power while simultaneously operating one of the most resource-intensive AI infrastructures on Earth. Training frontier AI models requires enormous amounts of computing power, energy, and funding, all of which remain concentrated among a very small number of companies.
That tension hangs over the entire announcement.
The post also suggests OpenAI sees governments becoming more involved in AI oversight as capabilities advance. The company again floated the idea of an international organization designed to coordinate frontier AI development and potentially slow progress if safety systems cannot keep pace.
That kind of language increasingly sounds less like a software company and more like a technology superpower preparing the public for what comes next.
Another notable detail is OpenAI’s vision of giving “everyone on Earth a personal AGI.” While the company does not fully explain what that means, the implication appears to be highly personalized AI systems capable of acting as assistants, collaborators, researchers, and decision-making tools for individuals.
Whether that future becomes empowering or deeply disruptive remains an open question.
OpenAI’s latest statement feels carefully designed to reassure regulators, investors, and the public all at once. The company repeatedly emphasizes safety, human control, and broad access. But beneath the optimistic language is a clear message: OpenAI believes AI capabilities are advancing rapidly enough that even AI research itself may soon become partially automated.
That is no longer science fiction rhetoric. It is now part of OpenAI’s roadmap.
Support independent tech journalism
NERDS.xyz is independently owned and operated. If you enjoy my coverage of Linux, AI, hardware, cybersecurity, and tech culture, consider supporting the site on Ko-fi.
Support NERDS.xyz