Generative AI – Short Notes
Practical notes for choosing and using AI tools with careful human review.
What it produces
- Generative AI creates new text, images, audio, video, or code in response to instructions or other inputs.
- It learns statistical patterns from data and generates outputs that resemble patterns it has learned.
- A generated result is not automatically a factual record or a reliable source.
Shape the result
- Describe the goal, audience, context, format, and constraints in the prompt.
- Provide suitable examples or reference material only when you have the right to share them.
- Refine the request in steps and compare drafts against the original goal.
Understand limitations
- Generated details may be fabricated, inconsistent, or missing even when wording sounds certain.
- Models may reflect biases in data and may struggle with niche facts or changing information.
- A tool’s capabilities, data use, and safeguards differ by provider and version.
Review before use
- Check facts, calculations, names, quotations, code, and image details before publishing or acting.
- Respect consent, privacy, copyright, and any disclosure rules that apply to the work.
- Use expert review for consequential or specialized decisions.