This paper describes a socio-technical study based on physical world scenarios of deceptive behaviour occurring in a virtual collaborative environment. An agent-based modelling (ABM) approach was adopted to visualise trustworthiness that can signal deceptive behaviour in virtual communications among social actors. The modelling strategies were guided by attribution theories toward an agent’s perceived trustworthiness. The assessment of an agent’s trustworthiness is based on their language actions as observable information-based behavioural cues, derived from objective semantic analysis. The consistency between an agent’s words and actions and distinctiveness of the agent’s behaviour when compared against his/her regular behaviour can serve as input data in each interaction. A set of ABM rules is proposed to systematically capture this interaction and assessment of trustworthiness. The logic behind a dyadic attribution can support computations on the trustworthiness of an agent.