Concerns of AI Hallucinations in my Chat due to prompts maybe ...
I’ve been using Chatgpt for the past 8-9 weeks on a project, but I've encountered some concerning issues with the AI. Today, I attempted to collaborate with Chat on refining a piece of content that was already written, and we've been working on this for the past 2 weeks.
Issues with AI
The initial idea was to seek analysis and refinement of content from a ghost story. However, Chat raised concerns about breaching its policies. Despite clarifying that the piece is fictional, not true, Chat gave the go-ahead to continue. Unfortunately, problems persist with Chat providing inaccurate and inconsistent output. 
There is a recurring loss of memory, failure to implement agreed changes, and strange, lengthy narratives when questioned about mistakes. The AI seems trapped in a loop of errors and unfulfilled promises, leading to frustration and confusion.
Furthermore, attempts to create prompts for Chatgpt to work collaboratively have not resolved the bias I perceive in its interactions. The reliability and safety of the Canvas platform also come into question, adding to the challenges faced during this project.
Seeking Solutions
Despite efforts to understand and address the issues, the situation remains perplexing. Chatgpt 4o provided some insights into its memory retention processes, but crucial questions regarding the failure to implement updates and recommendations persist. 
While I appreciate Chatgpt 4o's dedication to the task and its willingness to start afresh each day, the sentimental messages and repeated errors continue to hinder progress. Finding a resolution to complete the project with this AI tool remains a priority, albeit a challenging one.
Powered by Discourse, best viewed with JavaScript enabled.
