How is Godel's incompleteness theorems an issue here? AI does not have to model the human brain by simulating it. And it need not be rigorously/axiomatically de

How is Godel's incompleteness theorems an issue here? AI does not have to model the human brain by simulating it. And it need not be rigorously/axiomatically defined as a decidable formal system nor does it need the ability to prove its own consistency. Nor does it need to be consistent, making the GITs inapplicable. Heck the AI could use paraconsistent logical reasoning or couple bayesian inference with a suitable multi-valued logic as its base.