For those who are most enthusiastic about matters relating to fascinating topics related to Artificial Intelligence (AI), there are no great doubts today that what has motivated most opinion columns in recent times is unavoidably linked to the challenging topic related to ChatGPT ( Generative Pre-Training Transformer) and, in particular, with the supposed “new world” that it opens up to all of us, in the most varied and routine activities, whether in the personal orbit or in the professional sphere.
ChatGPT, launched in November 2022, is a system of chatbot free, AI-based and uses a language model developed by Elon Musk’s OpenAI.
As a main objective, ChatGPT seeks to simulate, in the most natural way possible, a conversation or any interaction between two human beings.
Powered by a complex and deep neural network, ChatGPT bases its operation on a self-learning model that is developed, above all, using the analysis of millions of texts, being able, in the end, to generate coherent texts with a smooth writing , semantically and conceptually intelligible.
Using a variety of applications, among which we would highlight virtual assistants, chatbots or even the generation of content, ChatGPT addresses, in essence, the efficient search for clear and coherent answers to the questions that are put to it.
However, as with everything in life, there is no beauty without a flaw, which is why we must also seek to question, from now on, among other relevant points, the nature and accuracy of the information that ChatGPT offers and the ethical implications that such a circumstance has. likely to behave.
Thus, and first of all, ChatGPT carries with it a high risk of bias or distortion, given that the information provided to its users is constructed based on a voluminous body of data.
It is enough to think of the fact that these data may be, to some extent, “contaminated” by any type of stereotypes, whether of a gender or even racial nature.
Given the model of linguistic nature on which it is supported, ChatGPT also poses a risk and a strong limitation in the clear understanding of the context from which the texts are generated, which naturally can lead the user to a poor interpretation of the response concrete that the system offers to a specific question that has been addressed to it.
On the other hand, ChatGPT still has limitations in answering questions for which, naturally, it does not have data, especially those that may be directly or indirectly related to events that occurred after 2021.
ChatGPT also contains some important and non-negligible ethical risks, given the potential use, with malicious purposes, that certain users may be tempted to develop, through the production and dissemination of false news that may endangermore or less intensely, social stability and peace.
To this extent, we cannot, nor should we, ever forget this relevant risk of misuse, malicious or even abusive use of this new technological system. On the contrary, and in a deliberate prudential strategy, we should rather seek to promote the creation of mechanisms that mitigate this risk and sanction, where necessary, all offenders and malicious persons in the use of this new and powerful technological tool.
One last note to leave some alerts regarding the need to guarantee privacy and security when using the tool. In reality, and considering that the tool’s linguistic model enhances the generation of knowledge based on the entries it receives, there is a non-negligible risk here of revealing sensitive information and content, and even at the limit, falling into the extreme of locate and profile specific individuals.
In summary, it will be important to create and have adequate measures that can effectively address the desired protection of the data of different users or prevent them from being misused and distorted.
Programmers, users and society in general will have to know how to live up to the different responsibilities that are now entrusted to them, in order to guarantee, smoothly, a successful use of this promising and auspicious technological tool.
This is the only way, we believe, that this same use will be effectively beneficial to Humanity, which is, after all, what the ChatGPT primarily aims to serve.
The author writes according to the new spelling agreement