The 14th ASEF Journalists Seminar (ASEFJS14) took place from 9-12 September 2024 in Budapest, Hungary, in partnership with the Danube Institute. As part of ASEF’s mission to foster media collaboration between Asia and Europe, 21 journalists were selected from over 500 applicants to attend the seminar.
Focusing on the theme “AI and Digital Journalism: Opportunities and Challenges,” the seminar featured five modules, covering topics such as audience analytics, AI’s impact on newsroom workflows and the automation of routine tasks. Participants delved into ethical concerns like AI’s influence on editorial decisions, accuracy and intellectual property. One module also explored AI’s potential for detecting misinformation and establishing guidelines for responsible AI use.
As the 14th ASEF Journalists Seminar came to a close, the diverse discussions and hands-on sessions underscored the transformative role AI is playing in journalism today. From enhancing newsroom workflows to tackling the ethical complexities of AI, participants gained valuable knowledge and practical tools to navigate this evolving landscape. ASEFJS14 offered a platform for meaningful exchanges between journalists from Asia and Europe on AI for journalism. Just as importantly, it provided an opportunity for young journalists from the two continents to meet, connect and hopefully, increase mutual understanding of their work and forge friendship.
We are very glad to share with you the takeaways from these modules, with its insights into and knowledge on AI and journalism.
Module 1: Using AI for audience analytics and content personalisation
By Pedro Henriques & Jenny Romano, The newsroom.ai
The module explored how artificial intelligence (AI) can be leveraged by media organisations to enhance audience engagement through data-driven insights and content personalisation. The session was divided into three key areas: understanding the value of user data and content metadata, transforming this data into actionable analytics and using these insights to offer personalised experiences. The first part emphasised the importance of mapping and enriching both user and content data, allowing organisations to better understand their readers. Examples from The Wall Street Journal and Reuters demonstrated how AI-driven metadata enrichment improved content discoverability and audience understanding. In the second and third parts, the focus shifted to transforming data into analytics for reader segmentation and using these insights for personalisation. AI was highlighted as a powerful tool for identifying audience segments based on shared characteristics and content consumption patterns. Read the report and presentation.
Module 2: AI’s impact on journalism workflow and productivity (newsroom backend) and on shaping media business models and journalism economics.
By Sannuta Raghu, Scroll.in
In this module the participants learned about how integrating AI into newsroom operations can enhance productivity and effectiveness. It defined AI as complex statistical software that identifies patterns in large datasets and discussed varying levels of automation, from basic decision support to full autonomy. Key use cases highlighted included news gathering, production, verification and distribution, all aimed at improving efficiency while maintaining accuracy and transparency. The module also encouraged journalists to build their own AI toolkits, focusing on practical applications that enhance their work. Participants learned about effective AI prompt elements—covering context, task definition, and constraints—alongside a list of useful AI tools designed to streamline daily tasks. Overall, the session emphasised the importance of both augmenting existing processes and rethinking workflows to adapt to the evolving landscape of digital journalism, ultimately empowering journalists to leverage AI effectively in their roles. Read the report and presentation.
Module 3: AI for automating routine news updates and stories
By Sannuta Raghu
The module provided an overview of AI’s evolution from the 1980’s to the most current development of AI and its application in journalism, focusing on automating news content. The agenda included a review of AI fundamentals, the automation process, and various use cases in the field. Participants explored different automation types, such as rules-based automation, multimodal LLM-based automation, automation with agents, and human-AI hybrid automation, emphasising how these methods can enhance content generation. However, the discussion also addressed significant challenges, including intellectual property concerns related to AI-generated content and potential reputational risks, as illustrated by recent complaints from The Times regarding inaccuracies produced by AI models. Further in the session, the module considered the human and environmental costs of AI, noting that the intensive training of AI systems has led to increased water usage in data centres, prompting a need for careful navigation of these issues as the industry evolves. Read the report and presentation.
Module 4: Developing guidelines and best practices for AI use in journalism
By Pedro Henriques & Jenny Romano, The newsroom.ai
The session explored AI’s integration in media, focusing on usage, challenges, risks and legal considerations. Participants learned the importance of AI in journalism, strategies for building guidelines and the best practices for adopting AI while mitigating risks. AI use in newsrooms is increasing, yet less than half of such organisations have formal AI guidelines. While tools like spellcheck and translation are widely accepted, readers are divided on AI’s role in generating content. Different regions have distinct regulatory approaches. News organisations should stay informed of evolving legal frameworks and ensure compliance with regional laws. AI poses risks, such as privacy concerns, bias and hallucinations. To mitigate these risks, organisations should adjust privacy settings, prioritise diverse datasets and employ human oversight. Read the report and presentation.
Module 5: Detecting misinformation and disinformation with AI (frontend)
By Dr Petra Aczél, Communication Specialist
This session raised probing questions about the inevitability of artificial intelligence (AI) in daily life and the inherent presence of falsehoods in communication. The discourse acknowledged that while some individuals may lag in adopting technology, misinformation pervades social exchanges, with humans often overestimating their ability to detect falsehoods. AI, though still influenced by its human-derived inputs, offers enhanced capabilities for recognising and managing misinformation through advanced data analysis and pattern recognition, which could potentially outpace human accuracy. Throughout the module, various aspects of misinformation and disinformation were explored and discussing AI’s role in identifying and managing them. Participants examined the philosophical and practical dimensions of AI in handling falsified information, using tools like the Rumsfeld-matrix to categorize knowledge and ignorance in this context. The session concluded by addressing the broader implications of misinformation in society, recognising it as a significant risk, potentially more impactful than war. Discussions included methods of falsification and the introduction of AI tools capable of identifying fake information through innovative techniques. Read the report and presentation.
Meet the trainers:
For a visual recap of the seminar, you can view photos from the event here.
