There is nothing to discuss.  AI is an engineering discipline.  It is not a branch of science or philosophy, but it can benefit from any results from any branch of philosophy or science.   Since metaphysics is a branch of philosophy and AI is disjoint from philosophy, AI is disjoint from metaphysics.  As for the ontology of AI, whenever AI engineers build something, an ontologist can assign it a place in any catalog of existing things.
However, philosophy can be used at the metalevel to study and analyze any subject whatever.   Since metaphysics is the metalevel analysis of any physical phenomena, it can be used to analyze any kind of physical system developed by AI engineers.
In any subject whatever, many of the practitioners study, analyze, and comment on the methods and results of other practitioners.  When they do that, they are doing metaphysics.  C. S. Peirce (1887) was one of the first metaphysicians who wrote about logic machines and what they can do.  Even earlier, Ada Lovelace wrote a metaphysical analysis of Babbage's machines.  Both Peirce and Lovelace were metaphysicians who contributed to the metaphysics of AI.  Alan Turing was another major contributor.  Marvin Minsky and John McCarthy were early contributors to the metaphysics of AI, and they continued their analyses until they died in the early 21st c.  Their methods of analysis and evaluation are just as useful today as they ever were.
That's all there is to say about the classification.  But there is a huge amount of work to do by metaphysicians of AI, who may also be engineers of AI.  But it's important to keep the two subjects distinct.   And the distinction between them was very clear to Peirce and Lovelace as it was Turing, Minsky, and McCarthy.   There is no reason to blur that distinction.

From: "Azamat Abdoullaev" <>
Sent: Friday, February 4, 2022 6:17 AM
To: "ontolog-forum" <>
Subject: [ontolog-forum] Is AI Metaphysics/Real Ontology?

I thought it might be a lively topic to discuss.