Sam Altman: AGI Term Losing Meaning in High-Stakes AI Race
Sam Altman, CEO of OpenAI, suggests that the term “Artificial General Intelligence” (AGI) is losing its relevance amidst the rapid advancements in the AI landscape. Once a rallying cry among tech enthusiasts, Altman now views AGI as increasingly passé, noting in a recent CNBC interview that it’s “not a super useful term” when questioned about whether the company’s latest GPT-5 model brings the world closer to achieving it.
Traditionally, AGI has been loosely defined as an AI system capable of performing any intellectual task at or beyond human levels. For years, achieving this advanced capability, while ensuring it is safe and beneficial for all of humanity, has been the core mission of OpenAI. However, Altman, who previously hinted on multiple occasions that the Microsoft-backed startup was nearing AGI, has recently shifted his stance. He now appears to downplay AGI’s significance, instead emphasizing a different concept: Artificial Superintelligence (ASI).
The challenge with AGI, according to Altman, lies in the multitude of definitions employed by various companies and individuals. He cited one common interpretation—an AI capable of performing “a significant amount of the work in the world”—as problematic, primarily because the very nature of work is in constant flux. Ultimately, he believes, the specific label is less important than the continuous, exponential growth in model capabilities, which humanity will increasingly rely upon for a widening array of tasks.
Despite this evolving perspective, the pursuit of AGI has undeniably been a crucial factor in AI companies like OpenAI securing billions in funding and achieving staggering valuations. OpenAI itself, based on its most recent funding round, is valued at $300 billion and is reportedly preparing for a share sale that could push its valuation to $500 billion.
While AGI remains a significant long-term objective for OpenAI, Altman advocates for a revised approach to measuring progress. During a FinRegLab AI Symposium last November, he explained, “We try now to use these different levels… rather than the binary of, ‘is it AGI or is it not?’ I think that became too coarse as we get closer.” He further predicted that AI-driven breakthroughs in complex fields such as mathematics and science could materialize within the next two years.
Earlier this month, OpenAI unveiled GPT-5, its latest large language model, which is now freely accessible to ChatGPT users worldwide. The company touts GPT-5 as smarter, faster, and more versatile, particularly in areas like writing, coding, and health-related queries. Yet, the launch has drawn some criticism, with a segment of observers suggesting that GPT-5 offers only marginal improvements over its predecessors, such as GPT-4o.
Responding to these assessments, Altman acknowledged at a media roundtable that GPT-5 is not yet at an AGI level. Nevertheless, he highlighted its profound impact: “The idea that you have a system that can answer almost any question, do some tasks, and write software for you at PhD levels of expertise… most people, if they heard that five years ago, would have said, ‘absolutely impossible.’” He concluded by emphasizing that the current impact of AI on sectors like education, healthcare, productivity, economic growth, and scientific discovery is “quite special.”