As multi-agent systems (MAS) become critical infrastructure managing healthcare, finance, and transportation, a fundamental paradox emerges: the very autonomy that makes MAS powerful also makes them untrustworthy. This paper presents six architectural design patterns that resolve this paradox by embedding trustworthiness into MAS foundations rather than treating it as an afterthought. These patterns, Transparent Decision Audit Trail, Consensus-Based Critical Decisions, Human-in-the-Loop Override, Graduated Trust Levels, Explainable Agent Behavior, and Fault Isolation and Recovery, transform abstract trust requirements into concrete architectural solutions. Through illustrative scenarios in finance, healthcare, and autonomous vehicles, this work demonstrates how systematic application of these patterns addresses the trust crisis hindering AI adoption. This position paper says that trustworthiness by design is not only desirable, but also crucial for the survival of society as MAS become its digital nervous system.