Not necessarily. Numerical integration of differential equations is function approximation. Yes the algorithm you would use may be an “artificial” implementatio

Not necessarily. Numerical integration of differential equations is function approximation. Yes the algorithm you would use may be an “artificial” implementation of “intelligence” but it is not AI in the sense people usually think of it today. The bigger picture is that function approximation is ubiquitous in applied mathematics, whether it be numerical calculus or statistics. Machine learning encompasses the state-of-the-art, big-compute techniques that have made statistical function approximation so much more useful.