Cross Fertilizing Empathy from Brain to Machine as a Value Alignment Strategy
This addresses the AI alignment problem for ensuring ethical AI actions, but it is incremental as it builds on existing literature and proposes future work without presenting new results.
The paper tackles the problem of aligning AI goals with human ethics by arguing that empathy is necessary, proposing a brain-based approach to algorithmically understand ethics and empathy, and concluding with suggested future research directions and initial observations.
AI Alignment research seeks to align human and AI goals to ensure independent actions by a machine are always ethical. This paper argues empathy is necessary for this task, despite being often neglected in favor of more deductive approaches. We offer an inside-out approach that grounds morality within the context of the brain as a basis for algorithmically understanding ethics and empathy. These arguments are justified via a survey of relevant literature. The paper concludes with a suggested experimental approach to future research and some initial experimental observations.