Given the increasing role of computerization and automation in organizations, computer algorithms are being increasingly used in marketing decision-making contexts. As is the case with any computer, it stands to reason that computer algorithms sometimes fail, which may result in brand harm crises. In this research, we examine consumers’ preferences for a brand following a brand harm crisis caused by humans (vs. algorithms). Extending developments in causal attribution theory and theory of blame, we propose that consumers’ preferences for a brand which faces a harm crisis caused by human error (vs. algorithmic error) will be more negative. We test our predictions in a series of studies and discuss the theoretical contributions and managerial implications of our findings.