Microsoft and OpenAI are moving toward a self-defined process for deciding whether artificial general intelligence, or AGI, has been achieved. The plan centers on an expert panel, but the most important parts of the process remain undisclosed.
That matters because AGI has long carried the weight of a scientific threshold. In this arrangement, it also becomes a contractual and strategic label controlled by the companies most directly affected by the outcome.
The decision will sit with a panel
According to the source article, Microsoft and OpenAI say they will appoint a panel of experts to determine when AGI has been reached. That may sound like an attempt to bring outside judgment into a difficult technical question.
But the companies have not said who will serve on the panel. They also have not explained what standards the experts will use, or what exact definition of AGI will guide the decision.
Those omissions are central. A panel can only provide clarity if the public understands its mandate, its membership, and the basis for its conclusions. Without those details, the process remains controlled by the same organizations whose business relationship depends on the outcome.
AGI still lacks a shared meaning
The source describes AGI as once being treated as a scientific milestone: an AI system able to think, learn, and solve problems like a human. That definition is broad, but it gives the term a clear direction. AGI is not simply better performance; it implies a more general form of capability.
In a joint podcast, Sam Altman and Satya Nadella made clear that there is no shared definition or timeline, even between them. That is a significant point. If the leaders of Microsoft and OpenAI do not publicly align on what AGI means, the final decision becomes harder to evaluate from the outside.
The uncertainty creates a basic problem for anyone trying to follow the technology. If AGI has no fixed criteria, then a claim that it has been achieved is not easy to test. It becomes a statement made inside a private framework rather than a milestone measured against a public standard.
Why the definition matters
Definitions shape incentives. If AGI is treated as a scientific term, it suggests a standard that should be explainable, repeatable, and open to scrutiny. If it is treated as a contract term between two companies, its meaning can become narrower, more flexible, and more dependent on business needs.
The source article argues that this shift turns AGI into a label that can be applied or withdrawn when convenient. That concern follows directly from the missing criteria. A term with no disclosed test can be difficult to challenge, even when it carries major consequences.
For readers trying to understand AI progress, the lesson is straightforward: the words matter less than the process behind them. A credible AGI determination would need more than an announcement. It would need a transparent explanation of what was measured, who evaluated it, and why the result met the threshold.
What remains unanswered
The current picture leaves several unresolved questions. They are not side issues. They determine whether the AGI decision will be viewed as a technical judgment, a business decision, or some combination of both.
- Who will be appointed to the expert panel?
- What qualifications or independence will those experts have?
- What definition of AGI will they use?
- What criteria will determine whether AGI has been achieved?
- Will Microsoft and OpenAI share a timeline, or avoid one entirely?
Until those answers are public, the process remains incomplete. Microsoft and OpenAI may be preparing a formal mechanism, but the mechanism has not yet provided the transparency needed to settle the debate.
A powerful label with uncertain boundaries
AGI has become one of the most important terms in AI, but this case shows how unstable the term can be. It can point to a scientific ambition, a public narrative, and a private contractual trigger at the same time.
That overlap is the source of the tension. If Microsoft and OpenAI decide for themselves when AGI has arrived, the decision may be valid inside their own arrangement. But for everyone else, the claim will remain difficult to assess unless the companies explain the standard behind it.
For now, the key fact is not that AGI is near or far. The source does not establish a timeline. The key fact is that the companies most closely involved in the outcome are also designing the process for naming it.