Artificial General Intelligence (AGI)
A hypothetical AI that could match or exceed human ability across the full range of intellectual tasks, rather than being good at just one. AGI does not exist today, and experts disagree sharply on whether or when it might.
Almost all the AI in use today is narrow: extremely capable at a specific job — recognising speech, generating images, playing a game — but unable to step outside it. Artificial general intelligence describes something different and, for now, imaginary: a single system with the broad, flexible intelligence of a person, able to learn and reason across any domain rather than one.
It is one of the most talked-about and most contested ideas in the field. Some researchers believe today’s rapid progress is a path towards AGI and that it could arrive within a couple of decades; others argue that current methods, however impressive, are missing something fundamental and that the goal is much further off, or not well defined enough to pin down. There is no agreed test for when it would have been reached.
The debate is not purely academic. Because a truly general intelligence could be enormously powerful, the possibility drives much of the work on AI alignment and safety — the effort to ensure that if ever more general systems are built, they remain under meaningful human control.