ChatGPT and Promotion of Artificial Intelligence in Sri Lanka?

International knowledge, based on Artificial Intelligence (AI), is the most important instrument in this process. 

by Dr Laksiri Fernando 

There cannot be any doubt that if Sri Lanka needs to get out of the present mess of economic fallout and political authoritarianism, solutions need to be worked out scientifically and objectively. This is something terribly lacking among the present political leaders and even among some of the young generations. In the case of the young people, however, many of the activists seem to be open minded and possibly work towards seeking new, innovative, objective, and scientific solutions. 

Credit: Alberto Romero via Midjourney

International knowledge, based on Artificial Intelligence (AI), is a most important instrument in this process. To expand one’s knowledge, intelligence, and awareness, new and innovative methods must be used going beyond the limited teaching in schools and universities, or teachers and academics. There should be a strong element of self-learning even among the teachers and academics.  

Chat GPT?

Have you come across Chat GPT? If you have not, it is simple to log on through ChatGPT: Optimizing Language Models for Dialogue (openai.com). Then you can chat and ask questions! But for what purposes? Objective knowledge should be the sole purpose. GPT would not answer inappropriate questions to my knowledge. I came across this device through my entrepreneurial son and he asked me to use this website to know more about the expanding world that I am seeking all along in my life.  

This Chat is a new setup in Artificial Intelligence (AI) developed by a multicultural expert group in America called OpenAI. You may also log on to the following OpenAI to know about them and there are frank and simplified explanations. In summary, the following is what they say: 

 “Artificial intelligence is an area of computer science that emphasizes the creation of intelligence machines that work and react like humans. Some of the activities computerized with artificial intelligence are designed to include: 

- Speech recognition, 

- Learning, 

- Planning, and

- Problem solving. 

The term ‘artificial intellegence’ was first coined by Dartmouth College professor John McCarthy in 1956. At the time McCarthy was one of the leaders of a small group of scientists who were exploring the possibility of creating intelligence machines.” 

The Chat GPT was created in November 2022 and not long ago. Still, it is in the making and there can be some problems or weaknesses. I have never been a computer person although I was one of the first to learn (or attempting to learn) computer science from Prof. V. K. Samaranayake at the Vidyodaya University in 1971. That was in theory. We had not seen an actual computer then. The program had to be temporarily abandoned due to the 1971 insurrection and consequences. The language we used was FORTRAN.  

How does it work?  

Today I managed to learn some about the Chat GPT, through nothing else but by chatting with the GPT! The responses to my questions were almost instant. Reporting some of my interactions, I believe, I might be able to promote some young people to utilize this mechanism to expand their knowledge, learning and problem solving. The final objective should be to modernize and develop the country and society. 

I asked, “what does GPT mean?” The answer was: “GPT stands for Generative Pretrained Transformer. It is a type of language model developed by OpenAI that uses deep learning to generate human-like text based on patterns it has learned from a large corpus of text data.” 

It is possible that OpenAI must have worked on the device based on the Internet for months or years before it was introduced to the world. But still there can be lacunae. Therefore, this device should not be considered perfect. They do not claim that even in answering questions. Then I asked the question “Why do you call ChatGPT a language model?” The answer was the following:  

“ChatGPT is referred to as a language model because it uses machine learning techniques to generate text that resembles human language. The model has been trained on a large corpus of text data and can be used to generate text responses to input prompts in a way that mimics human language patterns and styles. The ultimate goal of a language model like ChatGPT is to produce text that is coherent, relevant, and in context with the input prompt.” 

I believe the above answer is clear and, in the future, (if necessary) there can be a possibility to have different translations and or ask questions in a language you choose. But in my personal opinion, the world should more and more come together to work in a ‘lingua franca,’ as much as possible.  For the moment at least English serves the task. I then specifically asked the following question.   

“Would that mean the questions can be asked in different languages? Or translations can be done from another language to English or vice versa?” I raised the second part of the question, because at the University of Colombo, I was at one point advocating a computer (AI) program to instantly translate text or speech from Sinhala to Tamil and vice versa, and also to English and again vice versa. The GPT answer to my question was the following.   

“Yes, in theory, a language model like ChatGPT could be trained on data in multiple languages and used to generate text in those languages or translate between them. However, the quality of the generated text or translation would depend on the amount and quality of the training data in each language. OpenAI has trained a version of ChatGPT on multilingual data, which allows it to generate text in several languages, including English, but the model's performance may vary depending on the language and the complexity of the prompt.” 

On Some Subject Areas? 

It was interesting that the GPT started to chat as a person on some questions. Either this ‘person like’ responses were introduced newly or that is the way the machine was trained depending on the question. I asked the question: “Are you open in all areas of knowledge or specialized in some areas only? Following was the answer.   

“OpenAI has the potential to promote knowledge, intelligence, and research. OpenAI models, including ChatGPT, are trained on a large corpus of text data, which covers a wide range of topics and allows them to generate text that is informed by that data. However, the accuracy and quality of their responses may vary depending on the specific topic and context.

OpenAI is not specialized in some areas only, it is trained on a diverse range of texts, including but not limited to news articles, scientific papers, books, and online forums. This allows it to respond to a wide range of questions and generate text in various styles and formats. However, it's worth noting that the models are trained on text from the internet, which can include inaccuracies and biases, so their responses should be evaluated critically.”

I think the above answer is useful for any student, teacher, researcher, policy maker or administrator who would like to use this website. I continued and said, “My interests are related to democracy, human rights, and economic development. Is it ok for me to write articles based on your answers, but balancing them from my point of view or knowledge?” 

Then ‘he/she’ said the following which might be valid in general: “Yes, you may write articles based on my answers while balancing them with your own perspective and knowledge. However, it is important to note that I am a language model developed by OpenAI and my responses are generated based on patterns in the text I was trained on, which may not always reflect the most up-to-date information or nuanced perspectives.

It is recommended that you use my responses as a starting point for your research and supplement them with additional sources to ensure the accuracy and completeness of your articles. Additionally, it's important to ensure that your writing is original and does not plagiarize any sources, including my responses.” 

Then I also wanted to check the accuracy of historical knowledge that the GPT possesses! I asked, “Who was the author of the book Utopia published in the 16th century?”  The answer came: “The author of the book ‘Utopia’ published in the 16th century was Sir Thomas More.” Then I asked, “Author of the Prince? The answer came instantly, “The author of ‘The Prince’ is Niccolò Machiavelli.” This shows that at least the students of political thought can learn from GPT and not necessarily from a person Laksiri Fernando or others like him! 

Some Controversial Matters? 

I did ask some controversial questions. One was that “Do you think that war crimes were committed in Sri Lanka during the war between the government and the LTTE? Or by whom?” The answer was the following. 

“As an AI language model, I don't have opinions, but I can provide information based on the data I have been trained on.

There have been reports of human rights abuses and potential war crimes committed by both the Sri Lankan government and the Liberation Tigers of Tamil Eelam (LTTE) during the Sri Lankan Civil War, which lasted from 1983 to 2009. The conflict was characterized by widespread violence, including reports of extrajudicial killings, disappearances, torture, and sexual violence by both sides.

In recent years, several investigations, and reports, including those by the United Nations and human rights organizations, have documented evidence of human rights abuses and potential war crimes committed by both the Sri Lankan military and the LTTE during the conflict. However, the Sri Lankan government has faced criticism for failing to adequately address these abuses and for not holding those responsible accountable.

It is important to note that the situation is complex and that many of the allegations of human rights abuses and war crimes have yet to be fully investigated and proven in a court of law.” 

Let me end this article with the above answer from ChatGPT for the readers to make up their own minds on the usefulness and utility of this AI mechanism that I have talked about.