How to communicate with generative AI
Most generative Artificial Intelligence tools work through prompts: the instructions we give in order to guide the AI’s response.
Understanding how to formulate good prompts allows us to obtain more precise, useful, and academically relevant results for teaching, research, and university work.
What is a prompt
A prompt is an instruction or set of indications we give to a generative AI tool so that it carries out a task: explaining a concept, summarising a text, generating ideas, creating materials, or solving an activity.
The quality of the prompt directly influences the quality of the response.
Why a good prompt matters
AI generates its responses from patterns learned in large volumes of data.
The quality of the prompt affects learning, teaching, and research.
When a prompt is well formulated:
- The response is more accurate
- Contains fewer errors or hallucinations
- Aligns better with the goal
- Is more useful in an academic context
Limitations to keep in mind
AI tools cannot interpret hidden intentions or implicit information: they only respond to what is explicitly stated in the prompt.
Therefore, the clearer, more specific, and more contextualised the message, the better the response.
Additionally:
- Sensitive information should not be included
- Real personal data should not be included
- Nor copyrighted materials that cannot be shared
A prompt can be complete and detailed, but it should not be confused with a computer program: AI does not execute exact commands; it generates approximate outputs based on learned patterns.
Even with a well-crafted prompt, responses may vary, include unexpected nuances, or incorporate elements that were not initially intended.