Fishing for Real Needs: Reimagining Journalism Needs with AI
This video is featured in the Designing with AI 2025 playlist and 2 more.
Summary
Our team at Gazzetta, a media research lab, is tackling a fundamental challenge in journalism: the disconnect between media output and community needs, particularly in restricted or distorted information environments of autocracies. We have learnt over the past years that traditional audience research has led to quant-heavy, superficial understanding, ineffective content and, ultimately, irrelevance. To address this problem, we have developed a three-stage process using AI knowledge bases to build empathy, map information needs, and analyze information flows. We have used this process to systematically review multiple information sources to build deep community understanding before product development. This methodology has helped us preserve nuance, identify knowledge gaps, and assign confidence levels to findings. Rather than treating AI as a black box solution, a thoughtful process-oriented approach can help us better understand and serve information needs, and gradually rebuild relevance.
Key Insights
-
•
AI combined with structured research queries reveals deeper insights than traditional research when direct user access is limited.
-
•
Misalignment often exists between what organizations believe users need and what users actually need, as shown by North Korean fishermen preferring weather forecasts over political news.
-
•
Traditional engagement metrics only measure surface behavior, failing to capture true user information needs and satisfaction.
-
•
A shift from a theory of change (assuming known needs) to a theory of service (starting with understanding actual needs) is fundamental to effective media strategy and user-centered design.
-
•
A three-stage research approach—building empathy, prioritizing information gaps, and analyzing information flows—helps systematically leverage complex and scattered data.
-
•
Confidence rating of sources and insights is essential to prevent false certainty from AI-generated outputs.
-
•
Structured queries and systematic frameworks outperform opportunistic or freeform AI interactions, avoiding common pitfalls like overgeneralization and poor prompt design.
-
•
Human expertise, especially cultural knowledge, is indispensable to interpret AI outputs and maintain nuance.
-
•
Research repositories often hold untapped treasure troves of insights that, with structure and AI, can be mined even on shoestring budgets.
-
•
Visual coding of insights by confidence level improves transparency and decision-making among stakeholders.
Notable Quotes
"Journalists like me are in the business of interrogating reality to get at the truth."
"Those fishermen didn’t want political news—they wanted reliable weather forecasts to stay safe at sea."
"We spend millions on content that nobody wanted and that didn’t actually help people navigate their lives."
"Traditional metrics create a dangerous illusion where we optimize for what we can measure, not what people actually need."
"We’re moving from theory of change to theory of service: starting with what people actually need before creating anything."
"The careful application of structured queries can reveal deeper insights than traditional research alone."
"AI systems can present speculative connections as established facts, so confidence ratings are critical."
"Structure beats free form interaction—systematic query frameworks are essential to avoid pitfalls."
"Humans remain essential for evaluation, especially to interpret cultural nuance that AI often misses."
"Finding meaningful insights is not just casting a wide net—it requires discipline, structure, and knowing where to fish."
Or choose a question:
More Videos
"My kids know everything. If you listen to conversations with them, they say I know, Ma, I know."
Leisa ReicheltOpening Keynote: Operating in Context
November 7, 2018
"You can't just measure ethics like a scientific question; you have to reason differently."
Cennydd BowlesExit Interview #2: Rediscovering the ethical heart of design
November 6, 2025
"My playlist changed drastically when my son was born, but data alone wouldn’t explain why."
Andrew MichaelBuilding a Product Insights Team
March 10, 2022
"Since everyone was doing their own research, the quality of the research being done was pretty inconsistent."
Veevi RosensteinBuilding for Scale: Creating the Zendesk UX Research Practice
January 8, 2024
"Demographic criteria is generally considered less important compared to project-specific or contextual criteria in recruitment."
Megan CamposWhat Did I Miss? The Hidden Costs of Deprioritizing Diversity in User Research
March 12, 2021
"We realized designing for AI is very different from other computational mediums we used before."
Aras BilgenWho does the math: A designer’s journey in building an AI-based tutoring app
June 10, 2025
"It’s less about interaction and more about integrity: How can we build trust in experiences?"
Ron BronsonDesign, Consequences & Everyday Life
November 18, 2022
"Lots of people have agents running, but they don't feel they’re really useful yet."
Tina WeisserWhen AI Agents Meet Reality. Service Design Lessons from a Pilot
February 26, 2026
"Upskilling customer success managers and others to help with testing meant we could do far more research than I alone could manage."
Dr Chloe SharpUsing Evidence and Collaboration for Setting and Defending Priorities
November 29, 2023
