Amazon Just Fired the Human Voice Using AI. Your Job Is Next
Behind a simple Prime Video tweak is a glimpse at Amazon’s long-term AI content strategy.
Historically, Amazon has distinguished itself for its unrelenting focus on efficiency in service of improving margins and shareholder value. This was most recently demonstrated by its decision to lay off 27,000 people in 2022, and 30,000 people starting last month. But it won’t stop there. In the coming months and years, the company reportedly plans to replace roughly 500,000 jobs with robotic systems. Therefore, in the wake of massive cuts to its white-collar and blue-collar staff, it should come as no surprise that the company is aiming its automation filter at its creative unit as well, using AI.
This week, Amazon began rolling out AI-assisted recaps of some of its top original programming, including the hit video game TV series adaptation Fallout, The Rig, Tom Clancy’s Jack Ryan (starring John Krasinski of The Office), Bosch, and Upload. Previously, the company offered a recap button that provided text and slide-based recaps for various series. But that experience was something akin to going through a startup’s investor deck, efficient, but not something that necessarily inspired clicking the watch button.
In the updated version of Amazon’s recap feature, the company has deployed recaps that mix select scenes from a season along with a generic movie trailer voice reminiscent of Don LaFontaine, the late voice actor famous for his trailers that began with the words “In a world…” followed by the spoken plot points and central tension of the film. LaFontaine’s voice and his style, imitated by other movie trailer voice actors, became iconic and almost a part of the lore for some classic films.
Since the new tool is in beta testing mode, it isn’t available to all Amazon Prime Video users. Additionally, the AI-powered recap tool can currently only be sampled on traditional TV sets via the Prime Video app, not on the mobile app. When I tried it out, I only saw it as available for the Jack Ryan series. I’m not sure what I was expecting. Maybe a bit of AI-generated imagery, at least for interstitials or effects, or perhaps even a sample AI Gen scene. But no, all the scenes are simply footage from the show, no AI video. The only AI use appears to be the voice, and perhaps the selection of the scenes.
Reading Words vs. Breathing Life Into Them
I’ll admit, as a voice-over artist who has done professional work for companies like Nike, Snapple, and various organizations, I’m not happy to see the warmth and intent of voice actors replaced with AI. However, we should have seen this particular AI usage by Amazon coming. The company rolled out AI voice narration for its Audible audiobooks division earlier this year. Amazon touted the service by promising “100 AI-generated voices across English, Spanish, French, and Italian with multiple accent and dialect options.”
Superficially, to those less invested in the consumption and production of audiobooks, Audible’s initiative sounds fairly harmless. We already have several options for our computers to read the text on our screens. What’s the big difference in allowing an AI voice to do the same with a little more specificity and flair? In my view, the difference is the same in voice performances as it is with a screen actor’s performance: nuance and the human connection.
The first audiobook I ever experienced many years ago was J.R.R. Tolkien’s “The Silmarillion.” During a trip to the Netherlands, at a lakeside arts festival, I discovered the 1977 Caedmon Records vinyl record pressing featuring a map on its cover. Read in a commanding, stentorian voice by Christopher Tolkien, J.R.R.’s son, that audiobook brought more of Tolkien’s world to life for me in a way that I’d never experienced before.
“The art – and it is an art – of a good audiobook is the crack in the voice at a moment of unexpected emotion, the wryness of good comedy timing, or the disbelief a listener feels when one person can convincingly be a whole cast of characters,” Kristin Atherton, a veteran audiobook narrator, told The Guardian following Amazon’s announcement. “No matter how ‘human’ an AI voice sounds, it’s those little intricacies that turn a good book into an excellent one. AI can’t replicate that.”
2026: The Year Hollywood Decides If Humans Still Matter
Further complicating things for human voice actors is the rise of AI voice cloning, which allows companies like ElevenLabs to take a sample of a human voice to generate any audio you could ever want. Sure, the “acting” from that voice won’t necessarily be there, but for many famous voices, the familiarity will be enough for listeners to buy in and be satisfied. And aside from high-end paid AI services, there are already many free AI voice cloning services available online. So at least a significant part of the value of distinctive voices has already been erased, whether or not companies like Amazon’s Audible are involved.
I’m not sure any of this can be slowed down since there are so many fronts on the battle lines of generative AI: copyright vs. training data, human jobs being preemptively disappeared, and ongoing legal cases around protecting rights to human-created works. And while this multi-front AI evolution occurs, some companies, such as Warner, Sony, and Universal in the realm of AI music, are simply throwing in the towel and striking deals with AI companies to extract at least some value from their intellectual property.
As for Amazon’s tiptoeing into replacing actors with AI, I don’t really have a problem with an automated voice delivering a personalized weather update or letting me know the back door of my home is ajar. Most of that doesn’t require creative nuance from a human voice. And, admittedly, even this audio paired with real video AI recap didn’t hit my uncanny valley revulsion nerve as I thought it might.
But I also understand that this is baby steps for Amazon. The company has shown it will deploy machines to do any job that humans can do. That ethos will likely include creative outputs as well. Why would Amazon stop at AI voice recaps? Why not entire “live action” shows populated by AI actors? Personally, I’m excited about a future in which human actor-populated TV series and AI-“actor”-driven series might exist side-by-side.
My concern, and what does get my revulsion nerve twitching, is that once perfected, as with Audible’s AI narrators, the AI replacement of humans won’t be piecemeal. Instead, it will likely be broad, sweeping, and as unconcerned with the craft of human acting as Audible appears to be with the craft of human voice narration, which is absolutely a form of human acting. These people aren’t just “reading words.”
Anyone thinking that large Hollywood studios — similarly pressured by the expenses around hiring humans vs. AI automation — aren’t extremely interested in engaging in the same Amazon-style replacement of human actors on the visual rather than audio front, isn’t paying attention. This is the central tension leading up to the SAG-AFTRA negotiations for human actor rights slated to be settled (or not, which would result in another historic strike) in 2026. As LaFontaine might say: In a world… where art creation is offloaded to machines, what place does the human voice have? Better question: Will we even care?