Humans accept technology as a more impartial decision-maker in many environments, even when judgement involve moral decisions on other humans (although in some scenarios, human participation in the decisions is considered essential). Also, humans accept utilitarian justifications as suitable for implementation in software systems making moral decisions (despite utilitarianism is not widely accepted as the sole moral principle). Based on game theory, if machines are to make ethical decisions based on utilitarianism, while at the same time, machines themselves are not moral, optimal strategies would necessarily be mixed strategies. Thus, machines would be making randomised decisions. Randomisation challenges the long-standing notion that equates machines to deterministic artefacts. However, generative AI has propelled the acceptability of unexpected, divergent and randomised behaviour. We performed a study that shows that humans find it quite acceptable that machines could make moral decisions involving randomisation.