Watch: AGI (Artificial General Intelligence)
A system that matches or exceeds human performance across the full range of cognitive tasks rather than one narrow slice. There is no agreed definition and no accepted test, so progress is argued through benchmarks that stand in for the term.
Transcript
Artificial General Intelligence, or AGI, refers to a system that can handle the full range of cognitive tasks a person can, rather than just a narrow slice. Today, there is no official definition or accepted test for AGI. Every claim that we are close to reaching it depends entirely on who is defining it.
Some look at behavior, asking if a machine is indistinguishable from a human in conversation. Others define it economically, as a system that can outperform humans at most valuable work. Perhaps the most practical definition is labor-based: can a system hold down a job, learn on the task, and carry those lessons forward without needing its memory reloaded every morning?
Right now, we measure progress using benchmarks. But these tests are flawed. Test questions often leak into the training data, and scoring well on a test is not the same as being competent at a job.
This has led to a major disagreement. One group believes that scaling up current models with more data and computer power will get us there. Another group argues we are missing fundamental ingredients, most notably continual learning, which allows a system to improve while it runs.
Regardless of when AGI arrives, the practical landscape of artificial intelligence is already shifting. We are moving toward delegation. Instead of browsing search results, we are letting agents and assistants find answers for us. That transition is happening right now, no matter how we define AGI.