The term “Artificial Intelligence” has been bandied around for over 50 years to mean all sorts of things.
These days, all sorts of machine learning are generally classified as AI.
But I used to work with Cyc and expert systems back in the 90s, and those were considered AI back then, even though they often weren’t trying to mimic human thought.
For that matter, the use of Lisp in the 1970s to perform recursive logic was considered AI all by itself.
So while you may personally prefer a more restrictive definition, just as many were up in arms with “hacker” being co-opted to refer to people doing digital burglary, AI as the term is used by the English speaking world encompasses generative and diffusive creation models and also other less human-centric computing models that rely on machine learning principles.
According to gamedevs, 1-player pong (that is, vs computer) involves AI. It’s a description of role within the game world, not implementation, or indeed degree of intelligence, or amount of power. Could be a rabbit doing little more than running away scared, a general strategising, or a right-out god toying with the world, a story-telling AI. Key aspect though is reacting to and influence on the game itself or at least some sense of internal goals, agency, that set it apart from mere physics, it can’t just follow a blind script. The computer paddle in pong fits the bill: It reacts dynamically to the ball position, it wants to score points against the player, thus, AI. The ball is also simulated, possibly even using more complex maths than the paddle, but it doesn’t have that role of independent agent.
The term “Artificial Intelligence” has been bandied around for over 50 years to mean all sorts of things.
These days, all sorts of machine learning are generally classified as AI.
But I used to work with Cyc and expert systems back in the 90s, and those were considered AI back then, even though they often weren’t trying to mimic human thought.
For that matter, the use of Lisp in the 1970s to perform recursive logic was considered AI all by itself.
So while you may personally prefer a more restrictive definition, just as many were up in arms with “hacker” being co-opted to refer to people doing digital burglary, AI as the term is used by the English speaking world encompasses generative and diffusive creation models and also other less human-centric computing models that rely on machine learning principles.
According to gamedevs, 1-player pong (that is, vs computer) involves AI. It’s a description of role within the game world, not implementation, or indeed degree of intelligence, or amount of power. Could be a rabbit doing little more than running away scared, a general strategising, or a right-out god toying with the world, a story-telling AI. Key aspect though is reacting to and influence on the game itself or at least some sense of internal goals, agency, that set it apart from mere physics, it can’t just follow a blind script. The computer paddle in pong fits the bill: It reacts dynamically to the ball position, it wants to score points against the player, thus, AI. The ball is also simulated, possibly even using more complex maths than the paddle, but it doesn’t have that role of independent agent.