In the foreseeable future, autonomous vehicles will inevitably take control of our transport infrastructure. Although several challenges remain unsolved, researchers are already looking beyond these hurdles, preparing solutions for problems that lie ahead. One of the key motivations behind autonomous driving is the removal of direct human control. However, the absence of the human factor also introduces new issues. Concepts that are trivial to humans—such as trust, honesty, and fairness—are nonexistent in the strictly rule-governed world of machines. In this talk, we propose a possible way forward: reintroducing human-like behavior into autonomous systems—without the need for humans themselves. Predictive digital twin simulations can be used to assess trust between parties, while trust management techniques can help compute a complex trust score. Additionally, safety mechanisms can be adapted dynamically, based on the depth of trust, emulating human-like carefulness in uncertain situations.