Fine tuning the USER precedes the training of the device

ChatGPT currently functions as a substitute for a search knowledge engine with remarkable capabilities, marking a significant technological milestone in breaking down barriers and setting new standards for efficiency in information retrieval.

However, in my opinion, it is crucial not to rely (exclusively?) on the preferences of the majority of users for configuring (fine tuning) conversational devices.

Instead, emphasis should be placed on that specific segment of users who can formulate conclusive questions from the answers provided by artificial intelligence, using similar treats, logical thinking and hierarchical structure instead of personal biases.

For example, in an incorrect, off-topic, ambiguous, or misleading response, if the user asks why you responded in this way, they receive an apology for an answer, and the machine attempts to answer the previous question again! without providing information on the cause of the incorrect response. This is where the opportunity for successful machine configuration and effective fine tuning lies!

The goal should be to improve the conversational device to provide excellent answers ….NOT to perfectly formulated questions… but to help the user by asking questions about the initial question so that they can better formulate their own question, prioritizing quality over quantity of information.

So, user training in questions formulation precedes the training of the device, hence the need for a comprehensive list of “commands,” ensuring accurate answer results.

As technology evolves, conversational devices like virtual assistants and chatbots become popular, offering users convenient, hands-free assistance.
However, their accuracy in responding to users’ substantial questions remains imperfect (we are not talking about cooking recipes or where to go on vacation), so the dialogue itself and the formulation of questions by the user with the help of the machine should be the determining factor for more effective interactions and results.

The central logical configuration of conversational devices should prioritize and distinguish users who can formulate insightful questions from the answers produced by artificial intelligence, promoting excellence in question formulation instead of overloading with unnecessary information.

In conclusion, an option titled “Fine tuning the user” would be the ideal tool for “training” users and consequently the development and improvement of conversational devices.