TL;DR
Amazon has launched Alexa Podcasts, a feature that uses AI to generate entire podcast episodes on demand, narrated by two virtual co-hosts. The tool, available to Alexa+ subscribers including all Prime members, is backed by licensing deals with more than 200 news organisations including AP, Reuters, and the Washington Post
Amazon has launched a feature that uses artificial intelligence to generate entire podcast episodes on demand. Called Alexa Podcasts, the tool allows users to ask Alexa+ to create a podcast on any topic, and the system will research the subject, produce a structured overview, and deliver it as an audio episode narrated by two AI-generated co-hosts in a conversational format.
The feature began rolling out in the United States on 18 May 2026. It is available to Alexa+ subscribers, which includes all Amazon Prime members at no additional cost. Non-Prime users can access Alexa+ for $19.99 per month.
How it works
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
Users initiate a podcast by asking Alexa+ to create one on a subject of their choice, anything from a historical event to a scientific concept to a current news topic. The system researches the subject using available sources, generates a structured overview, and then allows the user to customise the episode’s length, tone, and focus before production. The finished episode is narrated by two virtual co-hosts whose voices are entirely AI-generated, delivering the content in the conversational back-and-forth style that has become standard in podcasting. Episodes are delivered as notifications on Echo Show devices and saved in the Alexa app for later listening.
The format is not new in concept. Google’s NotebookLM introduced a similar AI-generated audio overview feature in 2024, producing podcast-style conversations from uploaded documents. Amazon’s version differs in that it does not require users to provide source material, instead generating content from its own research, and in that it is integrated directly into a voice assistant ecosystem with more than 500 million Alexa-enabled devices worldwide.
The news partnerships
The most commercially significant element of Alexa Podcasts is not the on-demand topic generation but the news content pipeline that Amazon has built around it. The company has signed licensing agreements with more than 200 news organisations, including the Associated Press, Reuters, the Washington Post, Time, Forbes, Business Insider, Politico, USA Today, Condé Nast, Hearst, and Vox Media, along with more than 200 local newspapers across the United States.
These partnerships will power a separate capability that Amazon is also developing: personalised AI-generated news briefings that draw on licensed journalism to produce audio summaries tailored to each user’s interests. Media outlets have been experimenting with generative AI in newsrooms for years, from the Associated Press’s automated earnings reports to Reach’s AI-written local stories, but Amazon’s approach is different in kind: it is not using AI to help journalists produce content but to replace the act of reading or listening to journalism with an AI-synthesised alternative.
Amazon is also exploring the ability to generate podcast episodes from users’ own documents, a feature that would allow Alexa+ to turn uploaded PDFs, reports, or articles into audio content narrated by the same virtual co-hosts.
What it means for publishers
The licensing deals suggest that Amazon has learned from the mistakes of other technology companies that used publisher content without permission or compensation. But the structure of the arrangement raises its own questions. When a user asks Alexa+ for a news briefing and receives an AI-generated audio summary built from licensed journalism, the user has no reason to visit the publisher’s website, download its app, or subscribe to its newsletter. The content has been consumed, the informational need has been met, and the publisher’s only compensation is whatever Amazon agreed to pay in the licensing deal.
This is the same dynamic that has made AI-generated summaries a contentious issue in search, where AI Overviews have been correlated with significant declines in click-through rates to the websites whose content those summaries are built on. Amazon’s version is arguably more complete: a podcast episode is a self-contained product that does not even present the user with a link to click. The journalism that informs the episode is invisible to the listener.
The inclusion of more than 200 local newspapers is particularly notable. Local news organisations in the United States have been in financial crisis for more than a decade, with thousands of newsrooms closing as advertising revenue migrated to digital platforms. A licensing deal with Amazon provides immediate revenue, but it also risks accelerating the displacement of the audience relationship that local publishers depend on for subscriptions, donations, and community engagement.
The audience question
Amazon’s bet is that users want AI to do the work of finding, curating, and presenting information, and that the podcast format, with its conversational tone and passive listening experience, is the right delivery mechanism. The company is not the first to make this bet. Google’s NotebookLM, Spotify’s AI-generated playlists, and Apple’s personalised news digests all reflect the same thesis: that users prefer AI-curated content over the effort of selecting sources themselves.
But research consistently shows that most news readers say they do not want AI-generated content in their newsrooms, and the gap between what users say they want and what they actually consume when it is placed in front of them is one of the central tensions in AI-driven media. Amazon’s integration of the feature into Alexa+, which is bundled free with Prime, means that the barrier to trying it is effectively zero for the more than 200 million Prime members worldwide.
The broader pattern
Alexa Podcasts is part of a broader transformation of Alexa+ from a command-and-response voice assistant into an AI agent that produces and delivers content. Amazon has invested heavily in rebuilding Alexa around its Nova large language model, shifting from a system that answered questions to one that completes tasks, generates content, and acts on behalf of users across Amazon’s ecosystem of devices, services, and retail platforms.
The podcast feature sits alongside other recent Alexa+ capabilities, including AI-powered shopping recommendations, health information delivery, and smart home automation. Each of these represents a step toward Amazon’s vision of Alexa as the primary interface between users and information, a role that has historically been filled by search engines, news apps, and the open web.
For publishers, the calculus is familiar and uncomfortable. Major publishers have taken active steps to block AI companies from accessing their content without licensing agreements, and Amazon’s willingness to pay for access is a recognition that the content has value. But the long-term question is whether licensing fees can replace the audience relationships, advertising revenue, and subscription income that publishers lose when an AI intermediary stands between them and their readers, or in this case, their listeners.
