“Current OpenAI models are not considered AGI because they cannot continuously learn, improve on their weaknesses, or independently discover and integrate new scientific knowledge.”