We have proposed an Interactive Assistance system (IAS) to help the operators of a semi-automated industrial unit to increase their work productivity. At any moment the system suggests them to choose among a limited subset of possible tasks of higher priority. However, a remaining issue in this IAS is when one or more operators do not cooperate properly with it. We must take into account that they do not perform a task they have chosen and/or perform a task without indicating this action to the system. To address this issue, in this paper, we integrate a computational trust engine in the IAS. The IAS uses the trust engine to determine the trust level of an action carried out by a human operator.