Phil, I share many of your concerns, particularly the equivalence of AI and machine learning (further narrowing to deep learning). But I disagree with your assessment of the overall progress.
There are have been considerable advancements along the aspects of prediction and learning, important attributes of intelligence. I think the main question is the degree to which these functionalities reach across the full spectrum of intelligence.
As you’ve noted, the popularization of AI has masked these details. The problem is more systemic, as complex ideas are simplified for mass consumption. But the dominant success strategy isn’t evolving solutions as “artificial humans”, it’s an intelligent design, what you described as the “the framework of developing decision-making agents that solve problems”.