Last November 30, 2022, the company OpenAI made available to everyone an amazing technological breakthrough in artificial intelligence (AI). It is ChatGPT, Generative Pre-training Transformera highly advanced AI language model capable of answering almost any question in a coherent manner. Although still in the learning and development stage, the tool continues to be perfected on a daily basis and the advances it demonstrates are, to say the least, astonishing.
To use ChatGPT, you need to provide this artificial intelligence with a context question or phrase and, based on that, it will generate a unique and coherent answer with all the existing information in its data network. The only thing you have to do is to enter its web page and register if it is the first time you log in.
To learn more about this revolutionary technology, we talked to Dr. César Beltrán, an expert in artificial intelligence, professor in the Computer Engineering Section and director of Research and Innovation of our Engineering Department.
D. in Bioinformatics and Master in Computer Science, both at the University of Sao Paulo (USP), Brazil; he completed a post-doctoral stay at Texas A&M University (USA), 2016; he holds a degree in Systems Engineering from the Catholic University of Santa Maria (Arequipa), 1993. He is currently director of the Directorate of Research and Innovation of the Department of Engineering PUCP, has experience in artificial intelligence and has scientific publications indexed in journals.
What is artificial intelligence?
Artificial intelligence are algorithmic processes that seek to mimic the behavior of the human brain. With this, AI seeks to make decisions just as people do. Since Alan Turing (father of AI) sought to develop this type of systems that emulate human behavior, the Turing test was proposed, which is an examination of the ability of a machine to exhibit intelligent behavior similar to or indistinguishable from that of a human being; However, this law has already been surpassed some years ago, thanks to these chatbots. chatbots chatbots can engage a person in conversation by making it appear as if another person is answering.
In which fields is artificial intelligence already being used and in what way?
Artificial intelligence is considered the new revolution of our times, much like the discovery of electric power, which was one of the first revolutions that has changed our society. I believe that at this moment we as humanity should feel grateful because we are living a new revolution (such as artificial intelligence) that is changing the world.
Moreover, artificial intelligence is everywhere, as it is beginning to be applied, for example, in medical diagnosis, natural language processing, art, society, politics, and now even in science, such as AlphaFold, from the company DeepMind, which allows predicting the 3D structure of a protein from the amino acid sequence, a process that used to take months or years of work in biological laboratories. This company is also seeking to develop new algorithms that optimize matrix multiplication (AlphaTensor) and thus in other applications such as meteorology, nuclear fusion reactions, genetics, among others. We must consider that DeepMind started solving problems in games such as Go and Dota2.
Now, one of the applications that has recently taken the world by storm is the launch of ChatGPT. What is this?
The GPT Chat is a neural network trained with a large amount of information from the web. Its predecessors, GPT1 and GPT2, already showed hints that this technology could be developed to a practically human level and ChatGPT confirms this by processing natural language in the form of responses.
The nice thing about this tool is that, unlike its predecessors, its answers are much more coherent and not just structured like GPT 1 and GPT 2, which work on a single sentence and, in some cases, meaningless sentences. It is important to mention that GPT 4 is on the way, whose authors have only indicated that it will have greater capabilities than ChatGPT. I believe that in the next few years we will have a great revolution of tools of this type, which not only generate text, but also videos, music, among others.
There is a debate about the ethical use of this tool and its applications, what is your opinion?
I believe that it is just another technological tool and we must adapt to it, just as it happened when the Internet was born and practically did away with printed encyclopedias: before they were essential in the family library, today we look for any information we need on the web. ChatGPT is one more tool that will probably change the way we make queries on the web and also support us in making decisions. I believe that rather than fearing it, it is a tool that gives us new opportunities that we should exploit.
From an ethical point of view, I think it depends on who uses the tool, since like any technological tool, it can be used for good and for evil.
At the level of the PUCP, as a community, what benefits could this tool provide us?
I think it is a technology that can be used by both teachers and students. For teachers, it will provide us with a lot in the preparation of our materials and for students with a quick reference tool. However, all this will force us to change our teaching and evaluation strategies. It is no longer viable to continue with traditional strategies where teaching is centered on the teacher. Now more than before, teachers are the guides of the learning process. We should not be surprised that while we are lecturing on a topic in the classroom, students are consulting ChatGPT on the same topic, even more so in a virtual environment.
In what ways do you see the use of this technology becoming more widespread in our university?
In my last machine learning class in the Data Analytics diploma course, I used ChatGPT as an introductory tool to the topic. The topic of the day was about Bias and Variance in Ensemble models, so I took advantage of this to ask the chat about it. The answer he did not give left us surprised, as he coherently got the question right, making me start to develop my class in more depth afterwards. Therefore, I believe that we teachers can take advantage of this tool by taking all the potential it has and not be afraid of it.
"Artificial intelligence is everywhere, as it is beginning to be applied, for example, from medical diagnosis, through natural language processing, art, society, politics, and now even in science, through new structural discoveries that used to take months or years to be revealed."
Dr. César Beltrán
In your opinion, where do the advances in artificial intelligence point to?
The goal of those of us who research Artificial Intelligence is to provide machines with the ability to understand and learn any intellectual task that can be performed by a person, this is called General Artificial Intelligence. This may take 20 to 30 years, but we are on our way to it. As a society, we must create awareness about these advances, especially in countries like ours, because if we do not start to see the opportunities, the gap between countries could widen.
This leads us to reflect that as a country we are not creating the conditions to master this technology, and this involves creating the capacity of professionals in the field as well as building the appropriate infrastructure so that they can be used not only by the productive sectors of the country, but also by society itself (education, health, politics, among others). I am concerned that as a country we are not prepared to deal with it and, as with other technologies, we are simply consumers of it.
Comments