The Scoop
The Washington Post’s top standards editor Thursday decried “frustrating” errors in its new AI-generated personalized podcasts, whose launch has been met with distress by its journalists.
Earlier this week, the Post announced that it was rolling out personalized AI-generated podcasts for users of the paper’s mobile app. In a release, the paper said users will be able to choose preferred topics and AI hosts, and could “shape their own briefing, select their topics, set their lengths, pick their hosts and soon even ask questions using our Ask The Post AI technology.”
But less than 48 hours since the product was released, people within the Post have flagged what four sources described as multiple mistakes in personalized podcasts. The errors have ranged from relatively minor pronunciation gaffes to significant changes to story content, like misattributing or inventing quotes and inserting commentary, such as interpreting a source’s quotes as the paper’s position on an issue.
According to four people familiar with the situation, the errors have alarmed senior newsroom leaders who have acknowledged in an internal Slack channel that the product’s output is not living up to the paper’s standards. In a message to other WaPo staff shared with Semafor, head of standards Karen Pensiero wrote that the errors have been “frustrating for all of us.”
Another newsroom editor invoked recent criticism from the White House, underscoring how the paper has had to be cautious about errors and corrections.
“It is truly astonishing that this was allowed to go forward at all,” one Post editor wrote on Slack. “Never would I have imagined that the Washington Post would deliberately warp its own journalism and then push these errors out to our audience at scale. And just days after the White House put up a site dedicated to attacking journalists, most notably our own, including for stories with corrections or editors notes attached. If we were serious we would pull this tool immediately.”
One Post staffer told Semafor that although journalists have begun to use large language models for helpful tasks like transcription and research, lacking human oversight raised major red flags and opened the paper up to real concerns about editorial quality.
Know More
Internal frustrations with the Post’s AI-generated podcast reflect a wider series of tensions playing out within the paper around its audio products.
Over the last several years, Post leaders have batted around various concepts to improve the paper’s audio team, which some senior Post leaders feel has underperformed relative to competitors like The New York Times and The Wall Street Journal, which both have chart-topping podcasts. Semafor previously reported that the paper briefly was slated to release an opinion podcast co-hosted by former US Treasury Secretary Larry Summers (prior to public revelations about his closeness to Jeffrey Epstein), and former Post editor Sally Buzbee was engaged in other early-stage audio concepts before she was pushed out of the paper in 2024.
But Thursday’s tensions center largely on the differences in perspectives between the newsroom and its product divisions. The perspective from some among the Post’s product team has been that the AI podcast errors represent a normal part of the rollout of a feature that is still being tested.
Max’s view
The Washington Post is hardly alone in experimenting with AI news products. Yahoo announced earlier this week that it was rolling out almost the exact same product as the Post. Business Insider announced it was rolling out a tool for generating news stories overseen by human editors; the publication has published fewer than a dozen AI-written stories, one of which included a minor correction. Semafor has experimented with internal AI projects aimed at improving newsroom productivity.
Some outlets’ AI products have received backlash from news consumers, while others seem to have gone over without much protest. In an interview with Semafor’s Mixed Signals podcast, New Yorker editor David Remnick said the magazine’s readers have largely embraced the magazine’s AI-read audio longreads, which they can turn around much quicker than a host-read narrated article.
It remains to be seen whether a large number of people will begin to consume AI podcasts at a scale that can be profitable for media companies; at this point, it seems like the products are strange and buggy. (Eerily, the Post’s AI podcasts also include fake podcast tics, such as ums, uhs, and prolonged pauses, simulating the speech patterns of podcasters.)
But the rocky rollout represents a larger problem for the Post. The paper’s push into new products, including an opinion news aggregator launched this week, is trying to solve an audience problem with new widgets for which there’s little clear demand.
Since owner Jeff Bezos’ decision not to allow the paper to endorse Kamala Harris in the 2024 election triggered a massive backlash to the Post among its own subscribers, the paper has been attempting to reposition itself, making its opinion page more centrist and shedding some of its vehemently anti-Trump voices.
The ideological shift has driven its old subscribers, many of whom signed up for its antagonistic stance towards President Donald Trump in his first administration, into the arms of competitors like The Atlantic, The Bulwark, and The Guardian, who have poached the paper’s journalists and writers and seen a surge in subscribers that at times has coincided directly with negative news cycles about the Post.
Room for Disagreement
Tech and media companies have been aggressively rolling out AI podcast-like tools, many of which have excited different parts of the media industry. Speechify introduced a new AI audio summary tool earlier this year, while Google is using AI podcast hosts to answer search queries.
The View From The Post
The Post has been aggressive in 2025 in debuting several new products aimed at improving the company’s tech offerings and reaching new audiences, including Ask The Post AI, an in-app Watch tab, and Ripple, the paper’s news aggregator.
A Washington Post spokesperson declined to comment on internal discussions, but pointed to a quote from Bailey Kattleman, The Post’s head of product and design, in Digiday on Wednesday saying the personalized podcasts were “an experimental product in a lot of ways.”