Artificial intelligence (AI) is transforming the way we interact with technology, and one of the most powerful tools in this area is the use of prompts. A prompt is an instruction, question or request given to an AI model to generate an answer. It can be as simple as “writes a presentation email” or as complex as “writes a business proposal for a startup in the technological sector.”
“However, not all prompts generate useful or precise answers. Hence the importance of ‘Prompts Engineering’ arises, an optimization process that seeks to improve the interaction between humans and language models to obtain more precise and relevant responses, ”says Rogelio Mamani, a teacher at the Engineering career of Systems Engineering of Systems Systems Franz Tamayo University, Unifranz.
Specifically, explains the expert, Prompts engineering is an integral part of the use of AI, since it allows generating more natural and precise interactive conversations in chatbots such as Chatgpt, GPT-4 and Bard.
To ensure that AI responds effectively, the expert recommends five key tips for creating effective prompts.
1. Clarity and specificity: be direct in the application
One of the most common mistakes when interacting with AI is to ask too open or ambiguous questions. The clearer and specific the Prompt, the better the quality of the response that will be received will.
For example, instead of asking: “What do you think about robotics in medicine?”
It is more effective to ask: “What are the most significant advances in robotic surgery in the last five years?” This better guide approach to the model and avoids too general answers.
2. Adequate contextualization: Provide relevant information
Providing context is key when a precise response is needed, especially in complex or specialized issues. A well contextualized prompt helps the AI better understand the framework of the consultation.
If information about ethics is sought in education, the question could be asked as follows: “What are the most frequent ethical dilemmas in education and how have they evolved in recent years?” This type of prompt facilitates more focused and useful answers.
3. Coherent structure: Organize information
As in human communication, the way in which a prompt is structured influences the answer. If you want to obtain detailed information on several aspects of a topic, the question in clear parts must be organized.
For example, instead of asking: “Tell me about the history of artificial intelligence.”
Prompt could be structured in this way: “Explains the history of artificial intelligence by dividing it into three key periods: its beginnings, the era of neuronal networks and recent advances in language models.”
This type of formulation helps the answer to be more orderly and complete.
4. Use of practical examples to illustrate the application
When the issue is technical or abstract, provide concrete examples can make the AI better understand the application and deliver a more aligned response with what is expected.
If you want information about innovations in technology, you could write: “Describe three innovative technologies currently used.” This type of PROMPT allows to obtain more relevant and specific information.
5. Review and adjustment: Evaluate and improve the prompt
You will not always get the best response in the first attempt. If the result is not what was expected, the prompt must be reviewed and adjusted.
Mamani explains that when the questions are too general, AI can offer little precise or even diverted answers from the subject. However, when the PROMPT is clear, specific and uses relevant keywords, the answer will be much more accurate.
“Therefore, before making the consultation, you must check if the PROMPT is well formulated, free of errors and if you transmit exactly what you want to know,” he adds.
The success in interaction with AI models depends largely on the way the questions are asked. Prompts engineering is an increasingly valuable ability in a world where AI plays a fundamental role in work, education and digital communication.
