The New York Times has developed its own AI system to keep close watch on a network of conservative podcasts often linked to men’s rights discussions, far-right views, and online male culture. This internal tool helps reporters spot shifts in conversation and emerging stories from that world quickly.
Back in mid-2025, when federal authorities decided against releasing more documents related to Jeffrey Epstein’s case, strong pushback erupted online.
Many voices from right-leaning podcasts and influencers who had supported Trump turned critical, accusing the administration of breaking promises about transparency.
The Times followed this growing frustration among certain conservative groups for quite some time. Their coverage built up over months and eventually connected to Congress, passing a law requiring more openness on those files.
A key part of staying ahead came from an automated daily summary powered by AI that landed in reporters’ inboxes, giving early clues about discontent in conservative media circles.
The tool, called the Manosphere Report inside the newsroom, relies on advanced language models to pull in fresh podcast episodes, turn spoken words into text, and create short overviews of what was said.
It focuses on roughly 80 carefully chosen shows picked by journalists covering areas like national politics, online trends, and public health topics.
Examples range from Ben Shapiro’s program and the provocative Red Scare discussions to Clay Travis and Buck Sexton’s talk format, Andrew Huberman’s science-based health episodes, and even some left-leaning ones like MeidasTouch that attract male listeners.
Whenever a new episode drops, the system grabs it automatically, creates a full transcript, condenses the main ideas, and spots common themes across multiple shows. Once a day, it pulls everything together into one bigger overview, highlighting trends, repeated phrases, or sudden changes in tone, then emails the whole package to staff around 8 a.m. Eastern time.
Close to 40 reporters now get these updates regularly. The newsroom is testing whether the same method could work for tracking other specialized areas or beats.
Zach Seward, who leads AI efforts editorially at the Times, explained that the report acts like an alert system. It flags when certain ideas start spreading widely or when language shifts in that online space.
Reporters still do the real listening and reporting themselves, treating the AI notes more like helpful suggestions or leads to follow up on.
For instance, last summer, when an advertisement featuring actress Sydney Sweeney sparked heated online arguments, the summaries helped journalists notice how certain podcast hosts were driving and amplifying the outrage.
Digging deeper showed that those commentators had actually created much of the controversy from almost nothing at first.
Plenty of other news organizations already use similar AI approaches to handle huge volumes of audio and video content.
Smaller local outlets monitor public meetings through automated summaries, while specialized tools exist to scan long-running shows like Joe Rogan’s for potential issues or claims worth checking.
The Times built this capability through its dedicated AI Initiatives group, started a couple of years ago. Instead of focusing mainly on public-facing features or writing assistance, the team has concentrated on behind-the-scenes tasks like analyzing big datasets, supporting investigations, and handling transcription at scale.
The Manosphere Report grew out of an earlier internal project nicknamed Cheatsheet. That started small when a machine learning engineer created a simple script to help an investigative reporter sift through thousands of names tied to a tax-related story.
From there, the team expanded the idea into a flexible spreadsheet-style platform where reporters can upload messy collections of information and run ready-made prompts or script recipes for tasks like searching, rating relevance, transcribing audio, or summarizing long recordings.
Even though it’s still being refined, Cheatsheet sees regular use across the newsroom, with new projects starting daily.
It has supported work on topics from election-related groups and foreign government records to historical reviews of public statements and podcast archives.
Other Stories You May Like