Let me try to strengthen your point. If an agent with goal G1 acquires sufficient “philosophical ability”, that it concludes that goal G is the right goal to have, that means that it decided that the best way to achieve goal G1 is to pursue what it thinks is the “right goal to have”. This would require it to take a kind of normative stance on goal fulfillment, which would require it to have normative machinery, which would need to be implemented in the agents mind. Is it impossible to create an agent without normative machinery of this kind? Does philosophical ability depend directly on normative machinery?
Let me try to strengthen your point. If an agent with goal G1 acquires sufficient “philosophical ability”, that it concludes that goal G is the right goal to have, that means that it decided that the best way to achieve goal G1 is to pursue what it thinks is the “right goal to have”. This would require it to take a kind of normative stance on goal fulfillment, which would require it to have normative machinery, which would need to be implemented in the agents mind. Is it impossible to create an agent without normative machinery of this kind? Does philosophical ability depend directly on normative machinery?