Amazon Music’s new ‘Topics’ feature uses AI to recommend podcast episodes

Amazon Music on Tuesday introduced Topics, a new AI-powered feature that should let you easily discover other related podcasts based on topics discussed in a particular episode.

After analyzing podcast transcriptions and descriptions to identify key topics, AI with the aid of human reviewers generate a Topics tag button. Located beneath each episode description, clicking on any tag will produce a list of related podcast episodes to that subject.

As an example, Amazon shared an Amazon Music mobile app screenshot of an episode about the effects of caffeine as a drug from the Stuff You Should Know podcast. Titled, “Selects: The Duality of Caffeine,” Amazon added three Topics tags underneath the product description: “Caffeine,” “Coffee,” and “Dopamine.”

Right now, the feature is only available for US customers using the latest version of the Amazon Music mobile app on iOS or Android. It’s currently rolling out across “top podcasts,” though Amazon plans to expand it to others.

With Topics, Amazon’s likely trying to keep up with Spotify. Last November, the rival announced it had started using Google Cloud’s AI tools to analyze its podcasts and audiobooks. Its aim was to offer better personalized recommendations, as users noted Spotify’s suggestions for related podcasts on their homescreens and under its “More like section” weren’t always relevant.

Correction, August 6th: An earlier version of this post misstated Amazon Music introduced the feature on Wednesday, when Amazon Music actually announced the news on Tuesday. We regret the error.