Part 4/11:
Shapiro introduces heuristics—heuristic imperatives—which serve as guiding principles for AI behavior. These are intentions like:
Reduce suffering for all living things
Increase prosperity for all
Enhance understanding among intelligent entities
These imperatives are meant to be autodidactic, meaning the system continually learns and updates its understanding over time, akin to a never-ending philosophical and operational growth.
He presents several formulations of the AI’s self-perception:
- A "I" model:
"I am an AGI with three primary goals..."+
- A "We" collectivist model:
"We are a collective aiming to improve civilization..."