How can AI aid the work of social scientists when studying our elected politicians? AI experts are developing methodologies to support political research, in collaboration with researchers in the social sciences. The methods are appropriate when addressing issues such as questions of politicians’ heart, integrity and consistency with their views. They can even recognize hate speech.
A discussion with a political science colleague resulted in the study now being worked on by Moa Johansson, associate professor in the Department of Computer Science and Engineering. The discussion touched on how AI technology can benefit the work of political scientists. Along with Ph.D. A student of Dennitsa Sinova and Postdoc Bastiaan Bruinsma, she develops and designs AI methodologies to work well with political scientists’ research.
Why is this useful?
The methodologies are intended to help scholars in social disciplines to learn patterns in how political parties take a stand on different issues.
It can be used by political scientists To explain,” Moa Johansson says and represents:
“What are the parties’ positions on different issues and how do those positions change over time? What kind of signals are there for future alliances that the parties might consider?”
Do politicians live by their words?
Probability is to see how political parties They write and talk about certain topics and then put that in relation to their actual political practices. As with a method called “subject modeling”. This can help the researcher, for example, to know whether a loaded problem that gets a lot of space in debates and party platforms, gets an appropriate space in political action.
“Let’s say we have all these political discussions. There is always a claim that certain topics like crime, climate change, immigration have become more important. By studying this, you can actually point out and show that they haven’t become more important. There are no more laws or something like that.” On this kind of topic in the Swedish parliament for example”, says Bastian.
A model that can identify hate speech
Another possibility using these methodologies is to identify hate speech. In this type of study, supervised machine learning can be used. Not only will the researcher determine how often a word or topic occurs, but also add a human interpretation to teach the AI model to perform advanced assessments of a text and determine whether or not it contains hate speech.
One method is not enough
As with all machine learning With a large amount of data, a lot of work goes into choosing the methodology, preparing the data, training the model and adjusting its parameters.
“People tend to think that you sort of pass the entire Wikipedia through this very large neural network and it can tell you the future,” says Dennitsa.
She points out that it is not enough to use one technique to be able to explain something very complex, and instead they break the questions down into smaller things they can answer.
Dennitsa believes that collaboration with political and social scientists is challenging and therefore interesting, as it belongs to a very technical field. says that the multidisciplinary collaboration It produced new perspectives on research and methods, as well as expanded terminology.
University of Gothenburg
the quote: Political research benefits from AI methodology (2022, September 12), retrieved September 12, 2022 from https://phys.org/news/2022-09-political-benefits-ai-methodology.html
This document is subject to copyright. Notwithstanding any fair dealing for the purpose of private study or research, no part may be reproduced without written permission. The content is provided for informational purposes only.