PAREA is a good idea, gone bad. It had potential. But the leaders made a mess out of the idea. PAREA will only work if you get highly qualified college grads with a good set of mathematical/statistical, data mining, and programming. Give me that - and that is something to work with. And get some grads from tier-one university departments in math or computer science.
The entrenched bureaucracy does not want competition. In fact, the problem really is that top tier universities probably want to keep their distance from them. In fact, Traditional Appraisal and Academia is Oil and Water, especially in this day and age. They simply will not mix. They self-destruct when combined.
On the other hand, you can't just ditch TAF or the AI. We need to grow a new infrastructure in parallel and sit back and laugh as TAF and the appraisal organizations do their window dressing with AI-this and AI-that.
The only solution is to create a new parallel ( in the sense of existing side-by-side) system for what you might call "Valuation Engineering," based on strict protocols, mathematical constraints, and data mining.