This is just the Chinese room argument applied to the Chomsky vs Norvig debate, maybe with a dash of the hard problem of consciousness. Whether consciousness is inherent in particular structures or whether a statistical modeling can achieve it. Does the experience of experience deserve special pleading or would the simulation suffice?
https://norvig.com/chomsky.html
pull_request_target is criminally negligent -- github should simply disable it.
The security risk for running unvalidated code on any random PR with access to account secrets has no legitimate use case which outweighs its unbounded risk.