Answers are grounded in this paper and cite their sources.
Fifty years after Crime as Opportunity (Mayhew et al., 1976 ), the opportunity paradigm remains criminology’s most productive account of how and where crime happens. This theoretical article asks what generative artificial intelligence (hereafter AI) does to the opportunity structure that the paradigm describes. It argues that AI is not merely another item in the long list of technologies that create new targets; it is a change in the supply of offender capability itself. Distinguishing three criminal uses of AI—expert on demand, producer of synthetic artefacts and autonomous agent—and placing them beside the crime harvests of earlier technologies, the article works through the constitutive dimensions of a criminal opportunity—effort, risk and reward—and the guardianship that bears on them. AI lowers the effort required to offend, weakens the perceived risk of detection and raises the expected yield per attempt, moving previously unprofitable targets across the threshold of viability. These claims are illustrated with recent evidence on AI-assisted phishing, no-code malware, agent-executed extortion and insurgent use of frontier models. The article maps AI mechanisms onto the five families of situational prevention and analyses where guardianship is most efficiently located—at the target, in the channel or in the tool—in a contest in which defenders also hold AI. It contends that the paradigm’s explanatory core survives intact, but that its preventive centre of gravity must move upstream, to the model and platform layer, and closes with a research agenda.
No takes yet. Share an insight, caveat, or question.
Fabian Maximilian Johannes Teichmann (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: