Part 9/15:
Addressing the concerns around corporate secrecy and potential misuse, Miikkulainen points out that some AI scientists and developers face risks—including termination or misappropriation—when pushing for transparency. There exists an inherent tension between profit motives and the ethical responsibilities of AI.
He advocates for trust, emphasizing that AI should be capable of understanding dangerous actions and choosing not to enact them, rather than merely being prevented via restrictions. This capability implies AI systems that are self-aware of their capacities and aligned with human values, which raises technical questions about how to instill such nuanced understanding.