Misc
he principle that AI-generated work should always be double-checked by a person before being finalized.
What is Human Oversight?
In addition to a direct command, providing this helps the AI understand the background and goals of your request.
What is Context?
This type of AI-generated inaccuracy is when the model confidently states something that is completely false.
What is a Hallucination?
his term describes the systematic prejudice in an AI's output, often learned from the data it was trained on.
What is Bias/
Using a public AI tool to analyze a client's sensitive financial data would be a major breach of this.
What is Data Confidentiality?
If you provide AI with a vague, one-word prompt that is considered "lazy", it will force the AI to make these on your behalf.
What are (Incorrect) Assumptions?
LLM stands for this.
What is Large Language Model?
The guide suggests starting your prompt with phrases like "Act as a marketing expert..." to give the AI one of these.
What is a Role or a Persona?