Editors at The Washington Post are growing frustrated with the error-filled artificial intelligence podcasts that the paper has recently started producing. According to Semafor, earlier this week, the Post announced the rollout of personalized AI-generated podcasts for users of its mobile app.
In a press release, the paper explained that users would be able to customize their briefings by choosing preferred topics and AI hosts. They 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.”
However, just two days after the feature’s release, four sources reported that the podcasts were already being flagged for numerous errors. The mistakes ranged from relatively minor pronunciation issues to significant alterations of story content. Some podcasts misattributed or even invented quotes, while others inserted commentary that interpreted a source’s words as the paper’s stance on an issue.
Adding to the unsettling nature of these AI-generated podcasts, the Post included fake podcast tics such as “ums,” “uhs,” and prolonged pauses, mimicking the speech patterns of human podcasters. This approach has raised concerns about the overall reliability and professionalism of the content.
This development comes at a time of transition and tension within The Washington Post. Billionaire owner Jeff Bezos, who had previously taken a hands-off approach to the paper’s operations, has recently begun to take a more active role in shaping editorial policies.
https://www.rawstory.com/washington-post-ai/
