Safety signals provide relief and have been found to reinforce instrumental avoidance behaviour in non-human animals. However, the available evidence in humans has been obtained with subjective measures and hence there are no objective demonstrations of this phenomenon in humans. Using human participants in an avoidance task, three experiments were conducted online to assess the reinforcing properties of safety signals and a fourth experiment was conducted in the laboratory. Participants were presented with a CS+ and a CS-, and they could avoid an aversive outcome during presentations of the CS+ by pressing their spacebar at a specific time. If successful, the aversive outcome (IAPS image or a loud noise in the laboratory) was not shown but instead a safety signal was. Participants were then tested – whilst on extinction - with two new ambiguous test CS’s. If during test participants avoided, one of the new stimuli produced the trained safety signal and the other was a control. In Experiments 1 and 4 the control was followed by no signal, In Experiment 2 the control was followed by a signal that differed in one dimension (colour) with the safety signal, and in Experiment 3 the control was followed by a signal that differed in two dimensions (shape and colour) from the safety signal. We observed that participants made more avoidance responses to the ambiguous test CS when followed by the trained signal in Experiments 1, 3 and 4. We conclude that the trained safety signal reinforced avoidance behaviour. However, in Experiment 2 there was no preference suggesting that when the trained signal and the control signal are similar, generalisation occurs. Overall, these results suggest that trained safety signals can reinforce avoidance behaviour in humans.