Illustration by Tag Hartman-Simkins/Futurism. Source: Getty Images
Earlier this week, Washington Post Announced that it will launch “personalized” AI powered podcasts that will let users choose their own AI host to relight them on topics of their choice.
And now for a completely surprising update: AI podcasts have become complete, flawed disasters.
Semaphore report that Less than 48 hours after launching, AI podcast sparks outrage WaPo’s The rank and file and editors alike were outraged when they caught AI-generated podcasts committing terrible journalistic sins like inventing quotes and misrepresenting information.
“It’s really surprising that this was allowed to go ahead,” said one. WaPo Editors raged on Slack. “I never imagined this Washington Post Will knowingly distort its journalism and then broadcast these errors to our audience at large.
The editor added, “If we were serious we would remove this tool immediately.”
The podcast’s errors are exactly what you’d expect from an AI model. There are some simple but notable cases of mispronunciation. But sometimes, according to SemaphoreAI podcast hosts will inevitably insert editorial commentary by mistaking a source’s quote as the newspaper’s position on the issue.
in slack messages recipient SituationOther employees criticized the AI feature.
“What are the boundaries in place to ensure accuracy in this podcast?” One asked.
“This is a complete disaster,” said another Situation“I think the newsroom is embarrassed,”
Karen Pensiero, the newspaper’s standards chief, wrote in an internal message shared with staff. Semaphore The situation was “disappointing for all of us”.
Readers are also noticing errors. Jane Rosenzweig, a writer who covers technology, complained on bluesky He WaPo’s The AI podcaster “announced that they would later discuss ‘whether or not people with intellectual disabilities should be executed’ without mentioning any context.”
according to SemaphoreThere is an important difference between newsroom and of post Product Division. The podcast’s product team views errors as a normal part of rolling out a new and still experimental feature. Apparently journalists consider this an insult to their profession.
The podcast was developed in collaboration with AI voice cloning company Eleven Labs, and represents the latest way the newspaper has incorporated AI technology under Jeff Bezos’ ownership. Post is already using AI to provide summaries of its stories, and plans to allow non-professional writers to submit articles written with AI. It also has a dedicated “Ask the Post AI” page to ask questions to the trained chatbot on its articles.
More on AI: Time Magazine deploys AI “Ask Me Anything” box that covers its actual journalism and can’t be turned off
