In this study, LLMs are being evaluated as a mental health tool to combat the inaccessibility of mental health care in a time when there’s a shortage of psychiatrists in the US. Specifically, LLMs (in the form of chatbots) are being assessed as a safe resource for psychiatric patients, who tend to exhibit atypicality in thought processes compared to the general population, to use. While LLMs seem promising, this commentary brings up a concern of chatbots supporting delusional beliefs due to an assumption that a member of the general populace is being communicated with. Further, it is difficult to create a chatbot that can target multiple groups at the same time, such as those with atypical depression versus those with typical depression. That’s why this commentary proposed the dynamic contextual certification (DCC), an oversight process for chatbots with multiple stages: pilot contextual evaluation, targeted expansion trials, conditional general use, and lastly, continuous monitoring and revision.