Artificial general intelligence (AGI) is characterized by its ability to:

Artificial general intelligence (AGI) is defined by its capability to understand, learn, and perform a wide variety of tasks similarly to a human. This encompasses not just the ability to execute specific tasks or functions, but to adapt and apply learned knowledge across different contexts and domains—mimicking human-like flexibility in problem-solving and cognitive abilities.

This distinction from other forms of intelligence, such as narrow AI, highlights the broader cognitive abilities of AGI. AGI can integrate experiences, learn from them, and generalize knowledge to various new situations, allowing it to tackle unfamiliar problems without needing extensive retraining or predefined instructions. This characteristic aligns closely with human intelligence, which is not limited to rote learning or specific skills but is instead defined by adaptability and versatility.

The other options, while they describe aspects of AI technologies, do not capture the core attribute of AGI. Performing specific tasks better than humans refers more to narrow AI, which excels in specific applications. Storing vast amounts of data is indicative of data-driven AI systems but does not imply intelligence or understanding. Operating without human supervision may apply to various automated systems or narrow AIs, but does not denote the broader learning and understanding capabilities that define AGI.