Blog of the Data Science for the Digital Society research group. Digital Society Innovation, Applied Artificial Intelligence, data analysis and smart living and business.

21 February 2023 | Posted by angela.tuduri

Talking to ChatGPT - Technology of the future

The technological revolution and artificial intelligence have just begun.

ChatGPT -although you've most likely already heard its name- is a natural language model created by the artificial intelligence company OpenAI.  

The American giant aims to develop advanced artificial intelligence in a safe and responsible manner. The company is working on the research and development of new machine learning techniques and algorithms, including the Transformer architecture, used in the ChatGPT model. 

Sounds familiar? Read on!  

OpenAI - the company of the future  

In addition to developing advanced artificial intelligence technologies, OpenAI is committed to outreach and education in the field of artificial intelligence.  

Founded in 2015 by major tech entrepreneurs -Elon Musk, computer scientist Greg Brockman, philanthropist Sam Altman- and other experts in the field of artificial intelligence.  

We trained this model using reinforcement learning from human feedback (RLHF), using the same methods as InstructGPT, but with slight differences in the data collection setup.”

- CahtGPT

The company also conducts research and publishes reports and articles on its advances and discoveries, with the aim of encouraging debate and discussion around artificial intelligence and its implications in the technological world.  

OpenAI works together with other companies and organizations to help them incorporate artificial intelligence effectively and responsibly into their business operations and strategies. 

ChatGPT - operations and architecture  

Trained on a large volume of data including text from the web, books and other sources, ChatGPT uses deep learning techniques, and an architecture called Transformer.  

The main goal of ChatGPT is to be able to have natural conversations with people. This means that you can understand and answer questions, perform tasks and provide relevant information on various topics.  

I can perform natural language processing tasks such as question answering, text generation, language translation, chatbot assistance, and content writing, among others.”

- CahtGPT

The model is designed to adapt to the questions and answers it receives over time, improving its responsiveness with each question.  

Learn how to use ChatGPT - Operations 

ChatGPT works by using a natural language processing technique called "language modeling". This model is designed to predict a series of words in a sequence of text.  

For example, if the model receives the sequence "My dog is very", it is likely to generate the following word as "active" or "playful" because these are common words that usually follow "My dog is very". 

When a user enters a question or a message in the chat, the model uses this input to generate an answer. Through the context of the question and its prior knowledge, it generates the most likely answer.  

The model is designed to be adaptive and improve its responsiveness over time as it learns from interactions with users.  

In short, ChatGPT's main goal is to enable natural and fluent conversations between people and machines. To achieve this, it was trained on a large amount of data and developed using deep learning techniques.  

Transformer architecture  

The Transformer architecture is a deep learning technique used to process natural language. Developed by Google in 2017, it has been used in many successful language models, such as the one we analyze today, ChatGPT.  

Based on a fundamental idea called "attention", it is a technique used in natural language processing that allows a model to focus on different parts of a sentence while processing it.  

Instead of processing a sentence word by word, attention allows the model to focus on the most important words for the task at hand.  

Benefits and limitations  

Chat, which came to light barely a year ago -with the boom of new technologies and artificial intelligence- has certain benefits and limitations that we must consider before using it or implementing it in a business model.  


  • Scalability: ChatGPT can interact with many users simultaneously, making it a perfect tool for use in customer service applications and services, among others.  

  • Efficiency: It can process and generate responses at a much faster rate than most people, making it ideal for use in applications that require fast and accurate responses.  

  • Availability: Available 24 hours a day, 7 days a week for use in applications and services that require it.  


  • Limited knowledge: although ChatGPT has been trained on a large amount of text data, it is still limited in its knowledge and ability to respond to complex or out-of-context questions or situations.  

  • Bias: like any artificial intelligence system, ChatGPT may be biased due to the training data used. Therefore, it may have a limited perspective.  

  • Risk of misunderstanding: despite their ability to understand and generate natural language, they may still have difficulty understanding the context and tone of a conversation, which can lead to misunderstandings and inconsistent responses.  

The interview - Talking to ChatGPT  

Although the arrival of ChatGPT has not left us indifferent, many of us wonder about its efficiency and usefulness in everyday life.  


Discover the course

Find out how the interview with ChatGPT went!  

What is ChatGPT and how does it work?  

ChatGPT is a natural language processing model that is trained on large amounts of text data to understand human language and generate coherent and useful responses. It works by decomposing input text into smaller units and then generates output text based on those units. 

How does the quality of the ChatGPT responses compare with the quality of people's responses?  

While ChatGPT can generate coherent and useful answers to many questions, it is not perfect and can make mistakes. In addition, ChatGPT does not have the same ability as humans to understand context, culture and domain-specific knowledge. Therefore, it is important to be aware of the limitations of ChatGPT and use it in combination with human intervention.  

In which areas can ChatGPT be used?  

ChatGPT can be used in a variety of areas, such as question answering, text generation, language translation and content writing. It can also be used in chatbot applications, virtual assistants and text analysis tools.  

How to train and improve ChatGPT?  

ChatGPT is trained by using a large amount of text data, which is used to adjust the weights and parameters of the model. As more data is added and parameters are adjusted, the model becomes more accurate and efficient in generating responses. In addition, OpenAI's research and development team is continually working to improve and update the model.  

How can ChatGPT be used in the future?  

ChatGPT has great potential to enhance and facilitate a variety of tasks related to natural language processing. In the future, it is expected that ChatGPT can be used in even more advanced applications, such as real-time translation, sentiment analysis and personalized content creation. ChatGPT is also expected to be integrated with other artificial intelligence technologies, such as machine learning and computer vision.  

What is the main advantage of ChatGPT compared to other natural language processing models?  

One of the main advantages of ChatGPT is its ability to generate text consistently and in a human-like style. This makes it useful for applications such as chatbots and virtual assistants, where the ability to generate a natural conversation is crucial.  

How can the quality of ChatGPT responses be measured?  

The quality of ChatGPT responses can be measured using natural language processing evaluation metrics, such as the BLEU or ROUGE metric. In addition, it is important that responses are evaluated by humans to ensure that they are relevant, useful and consistent.  

The future will be technological and digital  

As technology continues to improve, AI is having a greater impact on our lives, from automating everyday tasks to improving healthcare and solving complex social problems.  

ChatGPT is a clear and exciting example of how artificial intelligence and natural language, and data processing can transform the way we interact with technology and each other.  

As technology continues to evolve, advances in AI and data language processing are likely to enable us to solve more complex problems and create new opportunities.  

La Salle-URL becomes an institutional member of the Association Catalan d'Intel-ligència Artificial (ACIA)  

The ACIA is a non-profit association that promotes artificial intelligence in Catalan society and aims to encourage communication between people and organizations.  

Its involvement in digital transformation, data analysis and artificial intelligence make it a Spanish benchmark, driving related knowledge in the area.  

La Salle-URL is positioned as a leading institution in Catalonia with studies and research in artificial intelligence, big data, IoT... It also joins the group of universities, institutes, and Catalan entities that promote the development and dissemination in a highly demanded area.  

The Master's Degree in Data Science at La Salle-URL responds to this need, training professionals with skills and knowledge for the analysis and management of data: the Data Scientist. 


Discover the program


Add new comment

This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
2 + 9 =
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.