Listen Notes: New Podcasts Bounce Back In January As AI Removals Surge.

The new year bounce back returned in January, as creators debuted 14,712 new podcasts last month. The podcast search engine and database Listen Notes says that is 1,500 more than it rolled out in November as the year-end holiday season typically delays show launches to the new year. The January list of new shows is, however, much shorter than what debuted a year earlier when Listen Notes reported 20,407 shows launched. But there is no shortage of new content to consume. More than 2.1 million episodes have already been published by podcasters worldwide this year.

The number of shows active is considered by many the true measure of industry health, and by that metric the data shows a robust start for 2026. Listen Notes says 321,493 podcasts have already been active this year. That compares to 362,047 in all last year. That continues a trend seen in 2025, when the number of active shows reached record levels.

Listen Notes also reports 1,971 podcasts are now considered to be “dead.” Listen Notes considers a show to have died when the RSS feed is deleted, or its iTunes “Completed” tag is marked “yes” by the publisher.

Overall, the tally shows there are at least 3,716,643 podcasts that have been launched to date. And 191,501,753 episodes have been published by those shows.

The new year also began with the largest number of AI-generated shows that were removed from the Listen Notes database since it began scrubbing that content. The company says 5,611 podcasts created by AI were removed from its search engine during January. That is more than twice the number of AI-generated shows it yanked a year earlier.

In an explanation for how it decides on what to remove, Listen Notes says it leans into technology to seek out “authentic, human crafted podcasts, not AI-generated junk.” It has developed an open-source tool to detect machine-generated audio content, including fake podcasts created by systems like NotebookLM. It also relies on human moderators who review every new podcast submission to ensure that it’s genuine and not spam.

“Keeping spam out of the database is a constant battle, especially as AI-generated content grows more sophisticated,” it says in a blog post. The company says most of the shows it removes come primarily via NotebookLM.