And the reason for that is because AI is only about mastering one task. For example, how to win a game. Artificial General Intelligence (AGI), on the other hand is about linking and applying knowledge to other tasks. For example, AGI would be the equivalent to how a mouse has learned to outwit a cat’s tongue.
The difference was explained really well in The Guardian last week:
Most AIs are based on programs called neural networks that learn how to perform tasks, such as playing chess or poker, through countless rounds of trial and error. But once a neural network is trained to play chess, it can only learn another game later by overwriting its chess-playing skills. It suffers from what AI researchers call “catastrophic forgetting”.
PathNet is a project being run by Google’s DeepMind to tackle AGI. According to DeepMind, PathNet is a neural network algorithm that uses agents embedded in the neural network whose task is to discover which parts of the network to re-use for new tasks.
James Kirkpatrick at DeepMind explains in the article that AGI is still a way’s off:
“We know that sequential learning is important, but we haven’t got to the next stage yet, which is to demonstrate the kind of learning that humans and animals can do. That is still a way off. But we know that one thing that was considered to be a big block is not insurmountable.”
So rest assure, lawyers and accountants, your job is safe for the time being!by