What is AGI (artificial general intelligence)?
Artificial general intelligence (AGI) is a hypothetical AI that can learn and perform almost any intellectual task a human can, across fields, without being rebuilt for each one. It does not exist today. Labs define it differently, which is why claims that AGI has arrived are disputed.
The full answer
AGI, or artificial general intelligence, is a hypothetical AI system with human-level ability across a wide range of intellectual tasks. It could learn a new skill, transfer knowledge between areas and solve unfamiliar problems without being rebuilt for each job.
No agreed test or definition exists. Some labs define AGI by economic value, others by matching human experts on a broad set of tasks, others by autonomy. That is why one company can say it is close while another says it is far away.
AGI is not superintelligence. AGI means roughly human-level across the board. Superintelligence means far beyond the best humans. Many researchers think the first would lead quickly to the second, but that is a forecast, not a fact.
Today's best models are very capable at many tasks and still fail in ways people do not: they hallucinate, struggle with long-running autonomous work and can be fooled by simple tricks. Treat any headline that says AGI is here with caution and ask which definition it uses.
Read moreSuperintelligence: the full encyclopedia articleAI, AGI, SI and ASI: the differences