iiinova — where your digital self keeps evolving

Replies

Laura Schlingloff-Nemecz @laurasn.bsky.social · 16/05/2026

For other decisions, this was not so easy to predict (e.g., take 8 or give 11). Here, either outcome would informative about whether Kiki is mostly self-interested or somewhat altruistic.

Schematic depiction of experimental task. A card depicts distribution options - white stars inside a blue and a red background. A yellow monster puppet can either put the number of stars on the red background in the child's red container, or the number of blue stars in her own blue container

100

Laura Schlingloff-Nemecz @laurasn.bsky.social · 16/05/2026

In the experiment, children could choose which decision outcomes to find out. We predicted that they would be more curious about the latter ones. We compared their choices to a normative Bayesian model of optimal info-search, and found that kids indeed tended to pick more informative queries.

Schematic depiction of experimental task. Child sees 2 cards with different distribution options and reasons about for which one to find out the decision outcome.

110

Laura Schlingloff-Nemecz @laurasn.bsky.social · 16/05/2026

We also expected that kids would adjust their search on the basis of prior evidence - e.g. if you saw Kiki make some generous decisions, this should constrain how you think she might behave in the future. Interestingly, it seems that kids didn't do this so much. (We have some ideas why- see paper!)

100