Artificial intelligence has become a critical issue across the world.
Various industries attempt to incorporate technology to adapt and lead
the future. People’s daily life has changed dramatically as information
technology and the internet advance. There is digital evidence of
people left untouched in cyberspace because collecting and utilizing
the data require specific resources and techniques. This article explains
how measuring public opinion has changed and identifies possible
difficulties in incorporating emerging technology. Public opinion is a
crucial subject in political science. Researchers use a survey to measure
public opinion and seek to utilize new technology like artificial
intelligence. Sentiment analysis is a type of opinion mining that
extracts attitudes from texts. Previously the analysis used frequency
based on a dictionary and, in recent days, utilizes artificial intelligence
and pre-trained models like BERT. Online public opinion can be
measured through this analytical technique. However, there are
difficulties to handle when researchers use artificial intelligence. The
styles of online texts are unique. It is expensive to build a dataset for
training artificial intelligence. There are many other formats of data
available besides texts. In addition, the application of artificial
intelligence poses challenges. It does not allow researchers to understand an algorithm as they do not provide a model like a typical
statistical analysis. It also requires significant computing power, which
can be unavailable to individual researchers. Nevertheless, artificial
intelligence is already a part of our everyday life. It is left to political
scientists to utilize the technique to gain deeper insight into political
phenomena.
Keywords: Artificial intelligence, public opinion, sentiment analysis,
research methods