Adjust for Trust: Mitigating Trust-Induced Inappropriate Reliance on AI Assistance