Can algorithms teach us to be kind and compassionate?

‘Loving-kindness meditation is a Radical Act of Love [to] maintain our individual and collective sanity’, writes Jon Kabat-Zinn, who secularised stress-reduction mindfulness meditation in the West. Initially in-person, group-based and teacher-led, mindfulness courses were then recorded for automated delivery through highly scalable mindfulness apps, which now boast tens of millions of downloads.

Cultivation of kindness and compassion (K&C) is encouraged in mindfulness courses because this can both reduce stress and help communities. In teacher-led courses, K&C are thought to be ‘embodied’ qualities that the teacher implicitly transmits via interactions with students more than via meditation exercises. Apps lack these interactions and try to cultivate K&C only via meditation. But is this effective? K&C meditation can be psychologically and culturally challenging, so apps with no human support could cause harm.

This project utilises interdisciplinary methodology from feminist health humanities and public health to investigate how K&C are constructed through mindfulness apps, how users understand such training, what effects they have on everyday experiences with K&C, and what unforeseen and/or harmful effects may arise from the automated cultivation of K&C.

This project is funded by the Centre for Artificial Intelligence and Digital Ethics.

University of Melbourne contributors:

  • Dr Julieta Galante, Deputy Director, Contemplative Studies Centre
  • Associate Professor Ana Dragojlovic, Insight Fellow, Contemplative Studies Centre