Learning to Be Epistemic Altruists

Suppose an epistemic agent has multiple strategies available. One maximizes their expected accuracy, while another sacrifices their own expected accuracy to maximize the expected accuracy of the epistemic group. Call the latter strategy “altruistic.” The importance of epistemically altruistic agents has been highlighted in many areas of formal social epistemology. But how do some agents come to be epistemically altruistic? In this paper, I demonstrate, using a self-assembling model with simple multi-agent reinforcement learning, how a group of agents can learn to be epistemically altruistic. Counterintuitively, competitive incentives at the individual level lead to cooperative choices at the group level.