02 / GENERAL INTELLIGENCE

Artificial General
Intelligence.

The idea of a machine intelligence capable of learning, reasoning and adapting across a broad range of intellectual tasks rather than being limited to a narrow domain.

What is AGI?

Artificial General Intelligence, commonly called AGI, refers to a hypothetical form of artificial intelligence with broad and flexible intellectual capabilities across many different domains.

Unlike systems designed primarily for particular tasks, an AGI would be expected to transfer knowledge between different problems, learn new skills, reason about unfamiliar situations and adapt to changing objectives.

There is no universally accepted scientific definition or agreed test for AGI. The term therefore represents a family of ideas about what general-purpose machine intelligence could mean.

What would make intelligence
general?

01

Generalization

Applying knowledge and learned concepts to new problems and situations that were not encountered directly during training.

02

Reasoning

Combining information, evaluating alternatives, solving unfamiliar problems and constructing multi-step conclusions.

03

Learning

Acquiring new knowledge and skills while improving performance through experience and feedback.

04

Adaptation

Adjusting strategies when environments, objectives, constraints or available information change.

05

Planning

Decomposing complex objectives into sequences of actions and evaluating possible paths toward a goal.

06

Transfer

Moving useful concepts and strategies between different domains instead of treating every problem as an isolated task.

Capability is not the same
as generality.

Modern AI systems can demonstrate remarkable abilities across language, vision, coding, mathematics and many other domains.

However, broad capability alone does not establish that a system possesses the robust generalization, adaptation, autonomy and reliability often associated with AGI.

The boundary between advanced AI and AGI therefore remains an open question rather than a universally agreed scientific threshold.

?

The difficult question is not simply whether a machine can perform many tasks — but whether it can understand, learn and adapt across fundamentally new ones.

From AGI
toward ASI?

If increasingly general intelligence can exceed the boundaries of individual tasks, a further question emerges: what happens when machine intelligence substantially exceeds human intellectual capability?

Explore ASI ↗