Revision and AI
A False Good Idea
Today, artificial intelligence can generate content and produce an answer in just a few seconds. An answer can be perfectly worded, seem convincing... and yet be incorrect.
When writing our multiple-choice questions and, above all, when answering our users' questions, we have chosen not to replace humans with artificial intelligence.
This is particularly important in the fields we cover. A small error in an answer can lead to a misunderstanding of an important rule or concept.
We therefore prefer to take a few minutes to check an answer rather than immediately provide an answer that we are not certain about.
Regulations evolve
New texts are regularly introduced, certain rules are modified, and older information gradually becomes obsolete.
Artificial intelligence may know an enormous amount of information, but it does not guarantee that the answer it provides exactly corresponds to the regulations applicable today.
When you ask us a question concerning regulations or a multiple-choice question answer, we can therefore take the time to check the information and, when necessary, refer to official texts or reference documents.
We prefer to be a human platform, with answers that may sometimes be slower, but verified, personalized and accountable.
Not just a simple question database
We know the questions we have created, the explanations that accompany them, the difficulties encountered by our users and the issues that have been reported to us over the years.
When a candidate asks us: “Why is this answer incorrect?”, you are speaking with a real person. Sometimes the answer will not be immediate, but we can:
* understand your question precisely;
* take the context into account;
* verify information;
* recognize that a question needs to be corrected;
* modify an explanation if it is not clear enough;
* and above all, answer you as we would like to be informed ourselves.
We prefer to admit when we don't know
An AI will always give an answer. We sometimes prefer to tell you: “We will check.”
For us, this is not a weakness. On the contrary, it is a guarantee of honesty.
In the field of training and regulations, we believe it is preferable to acknowledge uncertainty rather than confidently provide an approximate answer.
Using AI to revise
You can of course use artificial intelligence to revise. But if you really want to make sure that an answer is correct, it is preferable to cross-check the information with several AIs and verify their sources.
The problem is that two artificial intelligences can sometimes provide different answers to the same question. They can also provide a very convincing but incorrect answer, or rely on outdated regulations.
You then have to compare the answers, search for official texts, check the dates and determine for yourself which information is reliable.
You can use AI to revise, but to obtain a reliable answer, you have to carry out verification work and waste time.
Our questions and explanations are designed specifically for your exam and regularly checked and updated. When you report a difficulty or a possible error to us, a real person takes the time to verify it and respond to you.
AI is not our enemy
We obviously do not consider artificial intelligence to be a bad technology.
It can be extremely useful for many tasks. In particular, it can help us work more efficiently, analyze data or carry out certain technical tasks.
But when it comes to training, we believe that humans still have an essential role to play.
Your comments help us improve our questions.
Your reports help us correct errors.
Your questions help us improve our explanations.
And your feedback after exams helps us improve our content.
This direct relationship with our users is an integral part of our approach.
~
Updated il y a ~ 4 days