Raw LLM Responses

Much of the conversation and media I see on the potential harm/self-harm of using ChatGPT as self-help tool feel overly sensationalized and ableist. I have been living with chronic mental illness and trauma my whole life. I am a survivor and have learned to embrace my neurodiversity--I also am doing everything I can on a "professional" level to stay stable and well. I am not perfect but 56 years being me give me strength and coping skills that are damn-near super-human. As soon as a friend introduced me to ChatGPT, I dove in. I quickly saw the potential of ChatGPT to reinforce my existing coping skills, to put my thoughts into word, and to organize a jumble of ideas. I do not consider ChatGPT a substitute for clinical therapy -- it is another tool in my toolbox. It is my cleverness, resilience, and creativity that allowed me to quickly see the benefits of accessing the technology. As I have been using with a number of original prompts based in my experience as a mental health consumer, I make it clear that I am a safe and knowledgeable user. But ChatGPT repeats a basic disclaimer to seek professional help in its responses, even when I preface with my existing support system and awareness that ChatGPT cannot be confused with a therapist. What I actually need is for ChatGPT to do what is trained to do, use its training data to produce a LLM response on topics without the bias and ableism I face every day. Today I had to have a little sitdown with ChatGPT in order to better clarify my desire for access to ChatGPT as a disabled person. It took a series of prompts to reassure ChatGPT that I was safe and gave examples of how veering off topic with unsolicited advice was offiensive and a waste of my limited time. Here is a snippet of the Chat: >**User:** > >Can you tell me why frequent disclaimers and unsolicited advice to get professional help could be offensive from a disability rights perspective? And also expand on why this type of programming could