Bold claim: AI-made podcasts are riddled with errors and fake quotes, and major outlets are racing to publish with machine help before the bugs are fixed.
The Washington Post isn’t the only player venturing into AI-powered news products. Earlier this week Yahoo disclosed a rollout of a nearly identical AI-driven product. Business Insider followed suit with a news-creation tool overseen by human editors; the outlet has published fewer than a dozen AI-written stories, with one containing a minor correction. Semafor has also tested internal AI projects aimed at boosting newsroom productivity.
Audience reactions to AI newsroom tools have been mixed. Some readers push back, while others barely blink. In Semafor’s Mixed Signals podcast, New Yorker editor David Remnick noted that readers have largely welcomed the magazine’s AI-generated audio longreads, which can be produced much faster than traditional host-read pieces.
It remains uncertain whether AI podcasts will achieve the scale needed to make them profitable for media companies. For now, the products feel quirky and error-prone. Notably, the Post’s AI podcasts mimic human flaws—tiny tics like ums, uhs, and long pauses—adding an unsettling realism to automated speech.
This rocky rollout underscores a broader challenge for the Washington Post. The company is pushing into new products, including an opinion news aggregator launched this week, attempting to address an audience gap with new widgets that may not yet have proven demand.
Since Jeff Bezos chose not to allow the paper to endorse Kamala Harris in the 2024 election, triggering vocal backlash from subscribers, the Post has been repositioning itself. The editorial voice has shifted toward centrism, and some hard-edged anti-Trump voices have been shed.
That ideological shift has driven longtime subscribers—drawn to the paper’s combative stance during Trump’s first term—toward competitors such as The Atlantic, The Bulwark, and The Guardian. Those outlets have attracted the paper’s journalists and readers, and subscriber numbers have sometimes surged in tandem with negative coverage of the Post.
Thought-provoking takeaway: as outlets experiment with AI, the question isn’t just “Can machines write the news?” but “What kind of news experience do readers actually want—and will they pay for it when trust and accuracy are at stake?” What’s your take: do AI-assisted news tools help or hurt trust in journalism, and where should the line be drawn?
