Overview of this campaign
Some of the most important voices in American politics now operate on podcasts. Hours of conversation every week shape how millions of people think about elections, policy and power. But the podcast world is opaque. There’s no database of shows, no searchable archive of what was said and by whom. In order to understand this growing world, reporters have to sit there—and listen. For hours.
To help map this previously unsurveyable territory, we built ORCA.
ORCA (“They swim in pods”) monitors podcast feeds and automatically downloads new episodes as they’re published. Each episode is transcribed using AI that identifies individual speakers, so reporters know not just what was said, but who said it.
Google Gemini analyzes each transcript, generating multiple summaries and identifying the topics discussed. Each week, the system creates briefings that aggregate what’s being talked about across groups of podcasts—”politics,” “AI” or any custom selection.
Reporters can use a chat interface to ask questions in plain English. A clustering system groups similar episodes together, helping reporters spot emerging narratives and track how ideas spread across the ecosystem.
ORCA opens up a world previously impossible to search. For the first time, WSJ reporters can:
-
Track where ideas come from. When an unusual phrase shows up in a politician’s social media post, reporters can search ORCA to find where it appeared, mapping how talking points move from fringe podcasts to mainstream political discourse.
-
Get briefed. Weekly AI-generated custom briefings surface what matters across hundreds of hours of content so reporters learn what's happening on their beat without having to hunt for it.
-
Find who said what. Because ORCA identifies speakers, reporters can search for every podcast appearance by a specific person—a powerful tool for finding quotes and tracking public statements over time.
-
Spot trends as they emerge. The system groups similar content together, making it easy to see when a topic is gaining momentum or when a new narrative is taking shape.
Results for this campaign
ORCA enabled investigative journalism that would be prohibitively time-intensive without AI. A sampling:
“How the Epstein Scandal Fractured Trump’s Relationship With MAGA” (July, 2025)
ORCA analyzed conservative podcasts to produce the definitive account of how Epstein consumed the MAGA movement. The tool revealed 3,123 episodes across 124 shows mentioned Epstein—and that mentions exploded more than eightfold in the three weeks after the DOJ memo saying there was no client list, despite Trump’s admonishment to drop the issue. Reporters tracked the precise moment Joe Rogan drew “a line in the sand,” when Alex Jones turned on the administration and how the MAGA coalition splintered over the topic.
“‘Girl, Take Your Crazy Pills!’ Antidepressants Recast as a Hot Lifestyle Accessory” (October, 2025)
ORCA helped trace how antidepressants transformed from stigmatized medication to “#lexaprogirly” lifestyle accessory across influencer podcasts. The tool surfaced the voices promoting SSRIs to young women—and later, the same influencers quietly admitting to emotional numbness, brain fog and sexual dysfunction they hadn’t disclosed to followers.
“Kash Patel’s ‘Effin Wild’ Ride as FBI Director” (November, 2025)
ORCA’s transcript search surfaced FBI Director Kash Patel’s podcast appearances—including telling Joe Rogan the job was “effin wild” and his 2023 criticism of his predecessor for using government jets for personal travel.
Responsibility
When ORCA summarizes a podcast or answers a question, it points to the exact moment in the transcript—and original audio—so reporters can verify before publishing. Nothing is asserted without attribution.
The tool is designed to do what AI does well: automate tedious work, surface patterns and summarize large volumes of content. Journalists apply editorial judgment, talk to sources, verify facts and decide what’s actually a story. ORCA helps find leads, reporters decide what matters.
ORCA doesn’t write articles. It helps find them. The stories bolstered by ORCA sometimes required months of additional reporting. The AI simply made possible aspects of the reporting that were previously out of reach.