According to The Verge, Spotify is launching a new AI feature called Recaps that summarizes what you’ve already heard in audiobooks without needing to rewind. The feature, available automatically in the iOS app for limited English-language audiobooks, requires listening to 15-20 minutes before generating summaries that update regularly as you progress. Spotify’s research director of LLMs Paul Bennett confirmed the company isn’t using audiobook content for LLM training or voice generation, maintaining that original works remain protected. True-crime author J.H. Markert called it “an audio bookmark that speaks” in Spotify’s blog, while the company says it’s “designed to help people finish the books they start.” The feature launches amid Spotify’s broader AI initiatives, including February’s partnership with ElevenLabs for AI-narrated audiobooks.
The audiobook memory gap
Here’s the thing about audiobooks – they’re incredibly convenient, but our brains don’t always retain plot details the same way they do with physical books. You might listen during your commute, while cooking dinner, or on a walk, and suddenly you can’t remember who that minor character was or what happened three chapters ago. Spotify’s Recaps directly addresses this pain point that probably every audiobook listener has experienced.
But is this really solving a problem or just creating another layer of AI dependency? The feature seems genuinely useful for complex novels or non-fiction where details matter, but I wonder if we’re losing something when we don’t have to mentally track storylines ourselves. That said, for people who listen to multiple books simultaneously or take long breaks between sessions, this could be the difference between finishing a book and abandoning it entirely.
Spotify’s AI balancing act
What’s interesting here is how carefully Spotify is positioning this AI feature. They’re making it very clear that they’re not training models on copyrighted book content or replicating narration. After the backlash they’ve faced over AI slop in music, they’re being extra cautious about how they introduce AI into other content categories.
And they’re right to be careful. The publishing industry is watching AI developments like a hawk, especially after various copyright lawsuits around AI training data. By explicitly stating they’re not using audiobooks for LLM training, Spotify is trying to avoid the same controversies that have plagued other AI companies. Basically, they’re walking a tightrope between innovation and respecting content creators’ rights.
Where this is headed
Looking at Spotify’s recent moves – the ElevenLabs partnership for AI narration in February, these new Recaps, their work on “responsible” AI music tools – it’s clear they’re betting big on AI across their entire platform. The question is whether they can implement these features without alienating creators or degrading the user experience.
I suspect we’ll see Recaps expand to more languages and platforms soon, and possibly even more sophisticated features like character relationship maps or key theme summaries. The technology could eventually become so integrated that we won’t even think about it – just like how we now take for granted that streaming services will remember where we left off in a TV series.
For now though, it’s a smart move that addresses a real user need without crossing ethical lines. And in today’s crowded streaming landscape, those thoughtful, user-focused features are exactly what can differentiate one platform from another.
