Jaemark Tordecilla, Nieman Fellow at Harvard University, discusses the potential of custom GPTs for newsrooms. He shares experiences creating COA Beta Assistant for summarizing audit reports. Topics include streamlining journalism processes, integrating knowledge bases, democratizing AI access, and navigating model limitations.
Read more
AI Summary
Highlights
AI Chapters
Episode notes
auto_awesome
Podcast summary created with Snipd AI
Quick takeaways
Custom GPTs streamline journalism workflows by summarizing reports and aiding data analysis without coding.
Access to AI tools like custom GPTs democratizes technology in newsrooms, enabling investigative reporting and content creation.
Deep dives
Custom GPTs Transforming Newsrooms
Custom GPTs, such as the ones discussed in the podcast, are revolutionizing newsrooms by streamlining workflows and enhancing journalistic capabilities. Real-world examples, like Modena deploying over 750 custom GPTs, showcase how this cutting-edge technology is being used to facilitate tasks such as generating clinical study reports, summarizing contracts, and providing quick answers on internal policies.
Enhancing Data Manipulation with Custom GPTs
Custom GPTs are not only useful for structuring unstructured data from PDF files but also for processing large datasets from CSV files without the need for coding expertise. By utilizing the code interpreter feature, journalists can query data and extract valuable information for news stories, enabling data-driven journalism without traditional coding requirements.
Democratizing Access to AI Tools for Newsrooms
The emergence of custom GPTs and their diverse capabilities, such as image recognition and linking with external apps, is democratizing access to AI tools for newsrooms. This accessibility empowers journalists to experiment with generative AI technologies, aiding them in tasks like investigative reporting, data analysis, and content creation without extensive technical knowledge.
Empowering Journalists Through Iterative Development
Journalists are embracing custom GPTs for various applications in their daily tasks, encouraging iterative development and sharing of AI-generated tools within newsrooms. By continuously refining custom GPTs based on feedback and experience, journalists can enhance their workflows, generate new story ideas, and improve productivity while maintaining the essential role of human oversight for accuracy and quality assurance.
Jaemark Tordecilla, Nieman Fellow at Harvard University, joins host Nikita Roy to discuss the potential of using custom GPTs for journalistic purposes. Jaemark is the former editor-in-chief and senior assistant vice president of News and Public Affairs at GMA News, the Philippines' leading digital news organization.
For the past year as a Nieman Fellow, Jaemark has been pushing the boundaries of what's possible with custom GPTs – personalized versions of ChatGPT that can be tailored with specific instructions, knowledge, and capabilities to serve a particular purpose or task.
In this episode, Jaemark shares his experiences creating custom GPTs like the COA Beta Assistant to summarize dense government audit reports. He dives into how custom GPTs can streamline processes across the journalism value chain - from data analysis and visualization to content creation and fact-checking.
The episode explores the ease of building these AI tools without coding, integrating custom knowledge bases, and leveraging advanced capabilities like image recognition.
Nikita and Jaemark also discuss real-world use cases, the power of democratizing access to AI for smaller newsrooms, and navigating the limitations of these models.