Before we can examine how technologies contribute to social harms, we need to address a more fundamental question: what do we mean by social harm and harm more generally? Social harm represents the organizing concept of zemiology, yet, pinning down what it entails has been a longstanding challenge. The concept of social harm emerged in response to criminology's narrow focus on criminalized harms and, from the outset, sought to widen the frame to include harmful but legal practices – particularly those rooted in state and corporate structures (Hillyard et al, 2004; Hillyard and Tombs, 2004, 2007; Cain and Howe, 2008; Canning and Tombs, 2021). Early proponents of zemiology stressed that harm extends beyond what is criminalized, and that many of the most serious harms in society – poverty, inequality, environmental degradation – are legal, systemic and preventable. Most definitions of social harm therefore emphasize its social origins: harms produced by institutions, structures or practices that are shaped by human decision- making, and which could be organized differently (Pemberton, 2016; Tombs, 2019). What unites most accounts is the belief that social harm is distinct from natural harm, in that it is rooted in changeable social arrangements and human agency. In this way, the concept has been used to spotlight the often- overlooked effects of capitalism, neoliberal governance and state violence (Cain and Howe, 2008; Pemberton, 2016; Boukli and Kotzé, 2018; Lloyd, 2018). Still, one might reasonably ask: if we develop a sufficiently robust and inclusive concept of harm, do we still need the category of social harm? It's a fair question. But as long as we continue to use the term – and it remains useful – we should ensure that our account of social harm is able to meet the standards we expect from any theory of harm.