Mystery AI Hype Theater 3000 Crunching the Numbers (with Decca Muldowney), 2025.10.20
Oct 30, 2025
Decca Muldowney, a journalist and web editor at the Distributed AI Research Institute, explores the critical intersection of AI and data journalism. She highlights the urgent need for journalists to differentiate between AI hype and credible research methods. The discussion touches on the implications of 'vibe coding' for reporting practices, the reliability of LLMs in journalistic tasks, and the dangers of using AI tools without rigorous method verification. Decca warns against devaluing journalistic skills while emphasizing the importance of method transparency.
AI Snips
Chapters
Books
Transcript
Episode notes
Legacy Data Methods Are Not New AI
- Journalists have been doing many so-called "AI" tasks for years using non-LLM tools like TF-IDF and regular expressions.
- Lumping legacy data methods into nebulous "AI" obscures real methods and creates false novelty.
Document Methods And Verification
- Journalists must foreground detailed methods when using modern tools because audiences now demand verification.
- Explain which tool you used, how you verified outputs, and note risks like data retention and subpoenas.
Cleaning Data Is Core Reporting Work
- Treat data cleaning and coding as core reporting work, not burdens to outsource to LLMs.
- Manual analysis often reveals biases and stories that autopilot tools will miss.
