Related Post
The clouds that were covering the atmosphere of the 2024 US presidential elections began to clear, making the scene appear clearer to many analysts, who began to analyze the result, especially for the losing camp (the Democrats) led by Kamala Harris.
After several articles in the field of technology using “A dashboard measuring media sentiment towards the US elections using artificial intelligence. It was developed by two researchers in the field of artificial intelligence in cooperation with journalists on Al Jazeera, which tried to explore what the media sees and tries to focus on. In this article, we want to shed light on the tool, the developing team, and the future of artificial intelligence tools in the media.
The beginning of the idea and its purpose?
After the explosion of the artificial intelligence revolution over the past two years, Al Jazeera Net management found that the use of artificial intelligence must go beyond just normal use, and this technology must begin to be introduced more into newsrooms.
Therefore, the administration found that the best solution is to create tools that help journalists create a coherent story using artificial intelligence and relying on reliable sources, so that Al Jazeera provides this tool with the sources, the tool analyzes them, and then the journalist connects the threads and writes the story. The task was assigned to a technology page.
To do this, we had to, on the technology page, seek the expertise of people who have the ability and experience to create an analysis tool that can help journalists in important events.
Dr. Fahd Jubeir, an associate professor in the Department of Computer Engineering at the University of Jordan and director of the team developing the sentiment analysis panel, says, “The idea began through an informal conversation with journalists at Al Jazeera. The goal was to showcase the usefulness of artificial intelligence in analyzing topics and the tone of coverage in news articles.” “.
He continues, “Sentiment analysis is a well-known artificial intelligence technique for extracting emotions in text. There are many algorithms for performing sentiment analysis, and the error rate varies based on the technique and the complexity of the text. For example, we often find that sarcasm in text poses a challenge for artificial intelligence-based sentiment analysis.” “.
The doctor adds, “In our work, we used the ChatGPT model to analyze and extract sentiment for news articles based on the article summary.”
Creating artificial intelligence tools in the field of journalism that take into account ethics
Jana Al-Musalmani – a researcher in the humanities and digital societies and head of the Scientific Committee of the Artificial Intelligence in Media Conference at Al Jazeera Network – says that developing artificial intelligence tools that take into account ethics is a very sensitive and thorny topic.
Therefore, when working with the US Elections Media Sentiment Panel team, there were pressing questions about fairness, accountability, and transparency in the processes on which these tools rely and how to identify unbiased sources to feed them into the model.
Jana sees this initiative – to develop a tool within Al Jazeera Net that uses artificial intelligence to analyze media sentiment on important issues – as a good opportunity, especially since one of the major ethical challenges lies in the high cost of this technology, which requires advanced infrastructure and specialized personnel. This may lead to a widening of the gap between large and small media institutions, which strengthens the dominance of large companies over the media industry and increases their ability to influence public opinion.
Jana adds that one of the most important reasons that prompted her to participate in this initiative is the problem of lack of data, as Arabic content on the Internet does not exceed 3% of the total available data, which makes training models in Arabic a complex task.
Jana says, “Having a tool such as the sentiment dashboard and developing it in the future to include more resources in Arabic is a contribution to the field of supporting Arabic sources that feed artificial intelligence models.”
Create more tools to help journalists create stories
But will the dashboard for measuring media sentiment end once the elections end? Ahmed Hamayel – who works as a researcher and developer in the field of artificial intelligence and is one of the members of the team responsible for developing the sentiment dashboard on Al Jazeera – says, “The sentiment dashboard for elections is the beginning, as they are focusing on developing practical solutions that highlight the capabilities of artificial intelligence in an influential and tangible way.”
Ahmed adds that one of the biggest challenges is achieving a balance between the accuracy of artificial intelligence models and the speed of their performance, especially in applications that require an immediate response such as smart systems in real time, and therefore it was not possible in this project to create a dashboard that relies for its measurement on what is published in real time, which is Also one of the reasons for the inability to link it to social media, which are challenges that can be accomplished and overcome if resources are available.
Ahmed believes that artificial intelligence opens new horizons in multiple fields. For example, this Al Jazeera tool, the “Media Sentiment Dashboard,” can be developed to serve the media sector, as it can be used to analyze huge amounts of media content quickly and accurately, which helps provide deeper insights into… Public opinion trends and sentiment analysis in the news on specific issues.
In addition, it can be developed to enhance the user experience by customizing content to match the interests of each individual, so the user receives news that interests him personally.
It can also be developed to be able to examine sources and accurately identify fake news, which enhances the quality and reliability of the content. It is now in its current form, and on the subject of elections, it was able to reveal the hidden links between events and analyze their impact, which enabled journalists to build exciting stories, such as the media’s focus on personality and the integrity of elections, based on the available data, and to provide a clear picture of the effects of global events on public opinion.
What is the future of Sentiment Analysis Dashboard?
Dr. Fahd believes that Al Jazeera Net now has what is considered a nucleus upon which it can be built, especially since, in addition to sentiment analysis, it is possible, using artificial intelligence, to build time series forecasting models to predict future trends (Future Trends) for media coverage based on historical analysis of online news articles. These trends include both topics and feelings.
Jana stresses the need to pay attention to and develop this type of technical tools to preserve the presence of the Arabic language in the digital space. This tool and other tools allow accurate analysis of local texts, which enhances understanding of cultural and cognitive contexts. In addition, the local development process enables stakeholders and experts to discover problems and ask the right questions, leading to innovative solutions.
On the other hand, abandoning these tools and relying entirely on technology developed in non-Arabic-speaking countries threatens the future of the language and weakens its digital presence. Therefore, investing in the development of artificial intelligence technologies in the Arabic language should be a strategic priority to preserve cultural and linguistic identity in the age of technology.
