![\"AiPodcaster\"](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/07\/1784715311_688_0x0.jpg\")
<\/p>\n<\/p>\n
According to the M.I.T. Technology Review, rather than building thousands of smaller buildings, tech giants are investing in massive “hyperscale” campuses, with some individual AI sites requiring up to 5 GW of electricity\u2014more than the peak demand of entire U.S. states like New Hampshire.<\/p>\n
Frank Racioppi<\/p>\n
Artificial Intelligence (AI) has already seeped into our society, culture, economy, and daily routines. AI decides who gets their health claim paid. AI help millions with their administrative work at their jobs. AI answers our phone calls when we have a customer service problem. Millions of students try to use AI to ease the burden of learning, or to cheat. <\/p>\n
It was inevitable that AI would infiltrate the audio industry. In this article, we will appraise where AI has already affected the audio industry \u2013 audiobooks, radio, and podcasting. We will also assess how the consumer evaluates AI in the audio industry.<\/p>\n
Let\u2019s start with audiobooks. Last week, Spoken<\/a>, the AI Audiobook Company\u2122, released an intersting, \u201cindependent\u201d research study conducted by Edison Research at SSRS<\/a> showing that Spoken\u2019s multi-cast narration outperforms conventional human narration in engagement, favorability, and perceived quality. The study focused on character-driven fiction \u2013 the most prevalent category in today\u2019s audiobook market \u2013 marking a pivotal moment for the use of AI in the publishing industry.<\/p>\n The study randomized over 1,000 adult U.S. fiction audiobook listeners into two blinded cohorts, each experiencing samples of a new sci-fi thriller as either a professional, single-narrator human production, or an unedited, one-click Spoken Multi-Cast\u2122 narration. As industry stakeholders increasingly explore AI narration to meet the surging demand for immersive storytelling, this research offers the first large-scale look at how consumers respond to AI technology integrating distinct voices for each speaking character, a format historically reserved for full-cast human productions.<\/p>\n \u201cUp to now, we\u2019d only measured consumers\u2019 opinions on the concept of AI narration,\u201d said Megan Lazovick, vice president of Edison Research. \u201cFor the first time, we were able to measure real-time reactions to excerpts from an actual audiobook. Would there be a difference in acceptance between AI narration and human? What about listeners\u2019 willingness to listen or purchase? Could they distinguish the AI version from human? What we found was a clear signal that quality matters, no matter how the narration is produced, and listeners are open to whatever improves their experience.\u201d<\/p>\n Spoken CEO Phil Marshall added that this data confirms what he\u2019s believed since founding Spoken: \u201cAs an audio-only reader myself, getting lost in the story is what matters most. At the end of the day, what readers want will drive decisions in the industry. High-quality, multi-cast narration is what readers want, and a less expensive, one-click solution is what authors and publishers want. With Spoken\u2019s unique, patent-pending approach to layering multi-character scenes, we are able to deliver on that promise with an immersive audiobook experience that readers will pay for.\u201d<\/p>\n Podcast consultant George Witt casts a skeptical eye on the study. “Of course, It is somewhat suspect that the study, while not conducted by an AI audiobook company, validates the business model of the company that funded the study. Didn\u2019t tobacco companies use this tactic in the 60s to convince people that smoking was not dangerous to their health?”<\/p>\n Actor Simon Vance narrates the text for an audiobook of the novel Rules of Murder by author Juliana Deering in his recording studio at his in Concord, Calif. on Tuesday, May 7, 2013. (Photo By Paul Chinn\/The San Francisco Chronicle via Getty Images)<\/p>\n San Francisco Chronicle via Getty ImagesAccording to the study, while only 31% of listeners said they\u2019d be likely to listen to an AI audiobook at the start of the survey, after hearing an excerpt, 65% \u2013 more than double \u2013 said they\u2019d be likely to listen to an audiobook narrated using Spoken Multi-CastTM.They can\u2019t tell the difference: 61% of those who heard the Spoken version thought it was human, while 65% thought the human was human.The study also concluded that quality and immersiveness are more important than cost and celebrity: The top factors increasing likelihood of listening to AI-narrated audiobooks were improved quality, greater immersion, and using multiple character voices, not hearing a celebrity voice.<\/p>\n These findings arrive at an important time for the audiobook industry. According to Grand View Researc<\/a>h, the global audiobook market has experienced double-digit growth for many years and is projected to surpass $35 billion by 2030. <\/p>\n The Survey Methodology was as such: More than 1,000 U.S. fiction audiobook listeners were randomly assigned audiobook excerpts using human narration or Spoken Multi-CastTM AI narration and were asked a series of follow-up questions. The data was weighted to match the demographics of audiobook consumers as established by The Infinite Dial 2026 by Edison Research at SSRS and SiriusXM Media. Each treatment group was also weighted to match each other by age, gender, ethnicity, and interest in Sci-Fi\/Fantasy audiobooks.<\/p>\n \u201cNarrators don\u2019t just read words; they sense and express the feelings beneath the words. AI can never do this job because it requires decades of experience in being a human being,\u201d counters Audiobook narrator April Doty in The Guardian. <\/p>\n Adam Verner in The Literary Hub notes that, \u201cWhen a book is narrated aloud, a double art form is achieved, the book is interpreted by the narrator, who acts as a lens for the listener. Just as there are many productions of Ibsen\u2019s A Doll House, and each is unique to the actors and theatre performing it, every recitation of a book is a singular performance.\u201d<\/p>\n Is your favorite radio DJ a human?<\/p>\n Artifical intelligence has infiltrated terrestrial radio. Artificial intelligence is actively used on radio stations to serve as DJs, host unmanned shifts, and automate broadcasting tasks. According to Live365<\/a>, tools like RadioGPT combine synthetic voices, scriptwriting, and music selection to create entirely new, AI-hosted shows based on trending local topics. Radio Stations are using AI to fill overnight shifts, weekends, or timeslots that would otherwise feature pretaped, voice-tracked, or unhosted music, allowing for real-time local updates without human staffing.<\/p>\n More disturbing, however, is voice cloning capability, which means companies can clone a human host’s voice to create an AI double when they’re not available in-person. For example, Portland’s Live 95.5 famously debuted “AI Ashley”: a cloned voice of a human host in 2023. While it’s great that Ashley’s voice can stay on-air while she’s traveling or asleep, it does mean an up-and-coming human host willing to fill in for her just lost a potential job.<\/p>\n The “Guaranteed Human” program is an initiative launched by iHeartMedia in late 2025 that pledges all on-air personalities, podcasts, and music played across their 850+ radio stations are 100% human-driven. iHeart explicitly bans playing AI-generated music that features synthetic vocalists pretending to be human.<\/p>\n FilmMagic<\/p>\n There has been pushback, even from large corporations. For example, iHeartMedia \u2013 which operates hundreds of stations nationwide — implemented a “Guaranteed Human” program, requiring human hosts to identify themselves on-air to ensure authentic human connection in response to listener preferences.<\/p>\n The future holds more opportunities for AI to infiltrate radio stations. Futuri Media\u2019s Radio GPT<\/a> is an AI tool that combines different technologies to \u201chost\u201d a radio show. Basically, it combines the aforementioned synthetic voice, DJ’ing, and scriptwriting capabilities of AI to create entirely new radio shows on a station.<\/p>\n Radio GPT uses GPT-4 technology to create scripts and digital content based on what’s trending in a local market. Even more \u2013 it can produce content for the station’s website and social media channels in real time.<\/p>\n Stations can select from a variety of synthetic AI voices for single-, duo-, or trio-hosted shows, or train the AI with their existing personalities\u2019 voices. Programming is available for individual dayparts, or RadioGPT can power the entire station itself.<\/p>\n Is your podcast host a human or a synthetic?<\/p>\n Podnews <\/a>recently asked: \u201cIs podcasting being flooded with AI slop?\u201d For new shows, it seems that normal, human-made podcasts are now in the minority, according to the Podcast Index. Visible on the Podcast Index\u2019s New Feeds Report<\/a>, at the time of writing it shows only 44.6% of new shows in the last 24 hours were \u201clikely legitimate\u201d, with 45.7% potentially produced by AI. <\/p>\n The Independent<\/a> reports that more than a third (35% to 39%) of all new podcasts uploaded to streaming platforms are now generated by artificial intelligence, according to industry data from the Podcast Index. In fact, on some recent peak days, AI-generated shows have actually outnumbered newly submitted human-made podcasts.<\/p>\n \uf0b7 The Podcast Index New Feeds Report<\/a> also shows that Inception Point AI released 325 new shows on Tuesday: almost one-in-five of all new shows on that day. Hosted by Spreaker and monitored by Podtrac, it\u2019s unclear how many Inception Point AI shows have been produced, since the Podcast Index appears to sporadically remove them from its index; we count more than 8,000 still there.<\/p>\n ForbesWill AI Podcast Hosts Destroy The Very Soul Of Broadcasting?By Bernard Marr<\/a><\/p>\n AI-generated podcasts can be made in just minutes using numerous free online tools, with each featuring synthetic voices from artificial hosts discussing a topic prompted by the creator. <\/p>\n There are even dedicated AI podcasting companies, such as Inception Point AI, which churn out thousands of podcasts each week. The firm\u2019s website features various \u201cAI personalities\u201d, including an English gardener named Nigel Thistledown, a financial advisor called Pennie Power, and celebrity gossip monger VV Steele.<\/p>\n In a January 2025 Forbes <\/a>article, Jason Saldanha, Chief Operating Officer at PRX, criticized the use of AI hosts by warning that “flooding the market with content to get the lowest level of engagement” is not a long-term strategy. He emphasized that the true power of podcasts lies in “the host-audience relationship,” as the most successful shows have a “one-to-one relationship with their audiences.”<\/p>\n Just this week, Semafor<\/a> reported that The Washington Post\u2019s top standards editor Thursday decried \u201cfrustrating\u201d errors in its new AI-generated personalized podcasts, whose launch has been met with distress by its journalists.<\/p>\n Earlier this week, the Post announced that it was rolling out personalized AI-generated podcasts for users of the paper\u2019s mobile app. In a release, the paper said users will be able to choose preferred topics and AI hosts, and could \u201cshape their own briefing, select their topics, set their lengths, pick their hosts and soon even ask questions using our Ask The Post AI technology.\u201d<\/p>\n One Washington Post staffer told Semafor that although journalists have begun to use large language models for helpful tasks like transcription and research, lacking human oversight raised major red flags and opened the paper up to real concerns about editorial quality.<\/p>\n Getty Images<\/p>\n But less than 48 hours since the product was released, people within the Post have flagged what four sources described as multiple mistakes in personalized podcasts. The errors have ranged from relatively minor pronunciation gaffes to significant changes to story content, like misattributing or inventing quotes and inserting commentary, such as interpreting a source\u2019s quotes as the paper\u2019s position on an issue.<\/p>\n Just as in radio, podcasting creators and distributors are placing guardrails on AI.<\/p>\n Podcast hosting company RSS.com has added an AI disclosure feature<\/a> for all its podcasters. For now, it uses the podcast:txt tag, to ensure compatibility; and has issued guidelines<\/a> on when to use it. The company suggests that it is the first to add an AI disclosure tag to podcasts. It\u2019s in place in the company\u2019s Podcasting Innovation \u25b8<\/a> show. <\/p>\n Apple Podcasts requires<\/a> that “Creators using AI to generate a material portion of the podcast\u2019s audio must prominently disclose this in the audio and metadata for each episode and\/or show.\u201d<\/p>\n Sam Sethi of TrueFans and Podnews Weekly Review sees AI as a potentially valuable tool when used properly. Sam explains: \u201cAI is making it possible to analyze podcast content at a much deeper level \u2014 not just what a show is about, but what\u2019s happening within it. Themes, topics, and intent signals can now be identified and mapped to relevant brands.This is where contextual targeting is going next. From broad categories and genres, to specific episodes, and increasingly to the moments inside conversations where intent is forming or being expressed.\u201d<\/p>\n Sam adds: \u201cBut AI isn\u2019t creating value here. It\u2019s revealing it.\u201d<\/p>\n Sam warns that AI also creates a world in which, \u201cWe are no longer sure what we are seeing is real or created by an AI program. In the wrong hands, AI can be used for all kinds of nefarious purposes.\u201d<\/p>\n Tom Webster of Sounds Profitable<\/a> said, “In the Podcast Landscape 2025, we threw in a question for Americans 18+ who had ever consumed a podcast (75% of the population). Rather than ask about trying a hypothetical new podcast, we asked how people would react if they learned that one of their favorite podcasts featured AI-generated voices in either the content or the advertising. Would they be more or less likely to continue listening to or watching that podcast?<\/p>\n Interpreting the granular level data from the study, Tom Webster says: \u201cThe increased reluctance to embrace AI-voiced podcasts amongst those with college educations may be less due to ignorance, and more to an informed wariness about the existential threat this technology has and will continue to have for an entire class of workers. To succeed with this kind of content will not only require excellent execution, but also a means to address this wariness. AI may be coming to podcasting, but no one can be made to like it.\u201d<\/p>\n According to the Podcast Exchange<\/a>, \u201cThe podcast host-audience relationship is the most critical asset for a show\u2019s success, driving deep listener loyalty, retention, and monetization potential. Because podcasting is consumed primarily as a one-to-one experience (usually through headphones), this intimate medium creates a unique environment where the host feels like a trusted friend or companion rather than just a broadcaster.\u201d<\/p>\n AI: the Devil or an angel?<\/p>\n Does anyone want AI to replace these famous voiceover performances? Tom Hanks as Sheriff Woody in Toy Story; Eddie Murphy as Donkey in Shrek; Robin Williams as Genie in Aladdin; and the thousands of talented voiceover actors who bring audiences such joy.<\/p>\n Copyright Nick Brown Photography: Melissa Thom – BRAVA<\/p>\n High Notes<\/a> is a podcast featuring conversations on the art and business of voice, hosted by voice actor and BRAVA CEO, Melissa Thom. From Mongolian throat singing to vocal health, accents, gaming, and more, High Notes uncovers the craft behind the business.<\/p>\n Season four of High Notes is in production right now, and Melissa tells us: \u201cWe\u2019re focusing a lot of our attention on how AI is affecting our industry.\u201d<\/p>\n