The Future of ChatGPT According to an ITB AI Expert

Oleh Adi Permana

Editor Vera Citra Utami

BANDUNG,—Recently, ChatGPT has been in the spotlight as one of the biggest breakthroughs in the field of artificial intelligence and natural language processing. The technology significantly improved the generative language model and has a natural conversational capability, opening the door to more advanced and connected human-machine interactions.

From the Informatics Research Group of ITB SEEI (School of Electrical Engineering and Informatics) and the Head of ITB Artificial Intelligence Center, Dr. Eng. Ayu Purwarianti, S.T., M.T. explained that ChatGPT currently can only accept questions and provide answers in the form of text. “This does not rule out the possibility of receiving questions and giving out answers in images, audio, and video in the future.” With the same principle as ChatGPT, other available tools allow users to input words and sentences to be given answers in the form of images.

"Nowadays, there are many image and video generators that use deep fakes, such as making a video of a political figure using AI. This video does not require said figure to be involved, which is dangerous as it would mean that video is an act of black campaigning the person.”

ChatGPT uses a multilingual pre-trained language model that is trained with multilingual text data from different languages. The model is designed to understand and generate text in multiple languages without requiring separate modeling for different languages.

According to Dr. Ayu, ChatGPT is very helpful in speeding up one’s work, and it will be unfortunate if people do not use it often. "In the academic world, we can be taught the effective and efficient handling of ChatGPT. The study field Prompt Engineering will be useful in the future from this.”

In contrast, the possibility of ChatGPT’s development for better performance raises the warning for people to be more careful in choosing and performing their job, that their jobs are not easily replaced by AI. For example, many assume that television broadcasters and human translators can now be easily replaced with AI. Nonetheless, all outputs from AI still need to be validated by its user; people need to be more creative in making their work.

Dr. Ayu reminded us that we cannot avoid the progress of ChatGPT or other AI technologies; hence, people must adapt and understand how ChatGPT and AI in general works for wiser practice. People also must be aware of the risks of AI and learn how to overcome them. ChatGPT, for instance, provides "prohibitive" answers that must be re-validated and cannot be used as a primary source. Following the correct ethics, people can take advantage of AI technology for a better life.

Reporter: Satria Octavianus Nababan (Informatics Engineering, 2021)
Photo: Freepik
Translator: Ruth Nathania (Environmental Engineering, 2019)

scan for download