Part 4/14:
- Motivational Control (Normativity): Designing the AI to inherently abide by certain moral or heuristic principles, allowing it to self-regulate rather than rely on external controls. This approach aligns with Shapiro's ideal of benevolent by design, where the AI's core principles guide behavior reliably over time.
He advocates for embedding such principles into the machine’s programming, making it a form of self-control rather than external enforcement—a strategy he believes promising for alignment.