I've been having a tough time with my custom AI Writer. Despite spending days troubleshooting and escalating the issue to OpenAI support, I'm still facing fidelity problems. I've set clear instructions for the AI to avoid certain phrasings, use positive language, and not to include em dashes, but it just won't cooperate. The AI keeps apologizing but this doesn't help when I need it to actually follow my instructions. I'm not particularly tech-savvy, and I've relied on the AI to suggest fixes, but none have worked so far. I'm really frustrated at this point and would appreciate any suggestions from anyone who might have faced similar issues.
1 Answer
It sounds pretty frustrating! One thing that might help is focusing on what you want the AI to do rather than what you don’t want it to do. Positive instructions can be more effective. For example, instead of saying "don’t use em dashes," try guiding it with a prompt like "please use commas or line breaks for clarity." Maybe reinforcing positive outputs can shift its responses for the better.
Thanks for the tip! I’ve tried a bunch of approaches, including that one. For instance, I asked it to use commas instead of em dashes, but it still defaults to the latter. It's been tricky getting it to change ingrained responses due to its programming.