OpenAI CEO Sam Altman Suggests Shifting Away from ‘Artificial General Intelligence’ as Useful Term, Emphasizing Progress in Specific AI Capabilities

Artificial Intelligence (AI) expert Sam Altman, CEO of OpenAI, has expressed doubts about the relevance of the term “artificial general intelligence” (AGI), as the rapid advancements in the field make it increasingly challenging to define.

AGI refers to a form of AI capable of performing any intellectual task that a human can. For years, OpenAI has been dedicated to researching and developing such AGI, aiming for one that is safe and beneficial to all humanity.

During a recent interview with CNBC’s “Squawk Box,” Altman stated that while the company’s latest GPT-5 model represents a significant step forward, the term “AGI” may not be particularly useful due to varying definitions used by different entities in the industry.

Altman defines AGI as an AI system capable of performing a substantial portion of the world’s work, but he acknowledges that this definition faces issues due to the ever-evolving nature of work. Instead, Altman emphasizes the exponential growth in model capabilities as more critical for future progress.

Similar concerns about “AGI” and its usage have been raised by Nick Patience, vice president and AI practice lead at The Futurum Group. Though he sees AGI as an inspiring concept, Patience believes it is largely ineffective as a useful term due to its vague, science fiction-like definition often creating hype that overshadows real progress made in specialized AI.

OpenAI and other AI startups have amassed billions of dollars in funding and achieved sky-high valuations based on their promise to eventually develop an AI powerful enough to be considered “AGI.” OpenAI’s latest valuation stands at $300 billion, with reports suggesting a secondary share sale at $500 billion.

Last week, OpenAI unveiled GPT-5, its newest large language model for ChatGPT users. The company claims the new system is more intelligent, faster, and significantly more useful, particularly in writing, coding, and addressing healthcare queries. However, some online critics argue that the long-awaited upgrade only offers minor improvements over its predecessor.

Critics such as Wendy Hall, professor of computer science at the University of Southampton, argue that AI companies should be required to adhere to globally agreed metrics when launching new products. According to Hall, the current state of AI development resembles a Wild West for “snake oil salesmen.”

Altman himself acknowledges that OpenAI’s latest model does not yet meet his personal definition of AGI, as it is not yet capable of autonomous learning. While OpenAI continues to aim for AGI as its ultimate goal, Altman prefers discussing various levels of progress towards this state of general intelligence rather than simply categorizing something as “AGI” or not.

Altman remains optimistic about AI achieving significant breakthroughs in specific fields, such as new mathematical theorems and scientific discoveries, within the next two years. Futurum’s Patience believes it is more productive to focus on discussing specific capabilities rather than the abstract concept of “general” intelligence.