Integrating Imperfect Machines And Unmindful Users: Assessing Human-Bot Hybrid Designs For Managing Discussions In Online Communities