Critical Mistakes to Avoid with ChatGPT – ALWAYS remember ChatGPT is NOT encrypted. Your info is often shining bright for the world to access when used.
Unless using specific privacy modes, assume all data entered into ChatGPT is stored and reviewed.
Inputting Sensitive Data: NEVER share personally identifiable information such:
Birthday
Social security
Signons
Passwords
Company contracts
Company or org. secrets, legal docs. or evaluations
New business concepts
Data that could be used to train models/members/other companies
Relying on “Truth” or Logic: ChatGPT is a language model, not a search engine; it can “hallucinate” incorrect information and struggle with complex logic. It can recall simple facts well, so your ability to make direct factual asks can work well, if fact-checked. Historically and creatively it can prompt new ideas but essentially you’re still the ultimate decision-making and need to critically think and rethink given information.
Overloading Instructions: Requesting more than five major changes at once can cause the model to forget earlier, critical instructions.
Using it for High-Stakes Decisions: Avoid using it for medical diagnosis, legal advice, or financial planning. Nor is it your therapist. Seek true professional assistance when needed.
“Set it and Forget it” Writing: Using raw, unedited AI output produces generic, easily identifiable content that lacks human nuance. It works well for outlines, idea generating, fact finding.
New competitive options are coming out, like Claude AI, that you might want to explore…