It has often been cited that diagnostic aiding technology which falls below 70% reliability is not useful, and will harm overall task performance. This reliability threshold is based on tasks humans are capable of performing unaided. However, future robotic teammates may be capable of acting and gathering information and helping build situation awareness in environments that are too difficult or too dangerous for humans. However, if initial reliability is low, there may be resistance to introducing the technology. The current study investigates the perception of reliable and unreliable diagnostic aiding automation (robots) in both benign and dangerous environments. Undergraduate participants read a description of an autonomous robotic teammate working in either benign or dangerous environments, sending diagnostic aiding information to a human teammate with either high (80%) or low (50%) reliability. Participants in the dangerous environment conditions reported more positive overall perception of, and a stronger willingness to work with a robot, even at very low (50%) reliability. Results suggest that in dangerous environmental conditions, people may perceive unreliable robots more positively and be more willing to work with them. Implications for the introduction of new diagnostic aiding technologies, as well as strategies to support SA under conditions of unreliable diagnostic aiding are discussed.