Using Stories to Teach Human Values to Artificial Agents

91 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Scalable agent alignment via reward modeling: a research direction (2018, background): “This is analogous to how humans can learn about bad outcomes through story-telling and imagination (Riedl & Harrison, 2016).”
AI Safety Gridworlds (2017, background): “…RL (Hadfield-Menell et al., 2016a), learning from human preferences (see e.g. Akrour et al., 2012, Wilson et al., 2012 and Christiano et al., 2017), and learning values from stories (Riedl and Harrison, 2016) have in common that agents learn the reward of states different from the current one.”
Value Alignment or Misalignment - What Will Keep Systems Accountable? (2017, background): “While this trend has itself elicited some alarmed critiques of AI bias and systemic blind spots, there have been efforts to train for ethical values through richer kinds of material (e.g., narrative (Riedl and Harrison 2016)).”