Humanity has long sought, and failed, to achieve immortality.
Now the AI era has a solution. Sort of.
Eternos is a startup that creates AI voice clones of people so their loved ones can continue to interact with them after they’ve died.
It’s one of the more far out uses of generative AI, a technology with wide-ranging applicability, from generating content to planning summer travel, scheduling home repairs and translating baby cries. And as AI rapidly evolves, these use cases will only multiply.Â
After 18 years focused on business applications as the CEO of conversational commerce platform LivePerson, Eternos founder Rob LoCascio wanted to give consumer services a shot. More specifically, he was thinking about the 76 million baby boomers in the US.
“It’s a large population and they’re thinking about longevity and health, and I believe they’re thinking about longevity also in their legacy and their knowledge to pass along,” LoCascio said.
That led him to legacy preservation — and Eternos was born in April.
Eternos works with clients to create what it calls a “digital legacy” and a “digital voice twin.”
Once the AI replica has been created, loved ones can log into a portal and type questions like, “What was your favorite childhood memory?” into an interface that looks a lot like a traditional chatbot such as ChatGPT or Gemini. In return, they receive both a written and a spoken response.
Eternos can display letters, photographs and videos in the user interface as the AI replica is talking, but the replica itself is static. You won’t see a video avatar of your loved one speaking to you.
To create a digital voice twin, Eternos records each client saying 300 training phrases like “I love you” or “Shut the door.” That captures what LoCascio calls “your emotional voice.”
The startup feeds this into an AI model to create the voice replica.
Then, Eternos attempts to capture the breadth of the client’s knowledge in a 150-question interview on topics like memories, philosophies and even recipes, life hacks and final messages. Clients can tailor their questions to their experiences — so, say, a musician can opt to focus additional content on music. Clients can add and/or change the content accessible to their AI twin at any time during their life.
After the voice and knowledge have been captured, loved ones can ask about personal experiences from the past, as well as for advice or to read a bedtime story.
“It does what we know of the AIs today where it can read a bedtime story and make it with characters and do all that,” LoCascio said. “So it’s got facts, it’s got reasoning, it’s got what I’ll call generative as part of the engines.”
Each AI replica is an individual model that’s trained on data from that person. But AI immortality doesn’t come cheap. For now, it’s a one-time fee of $15,000, although LoCascio said the price will eventually come down.
“Right now, it takes a lot of GPU power to make you a very high-end AI,” he said. “I wanted to keep the quality very high [rather] than put something out that’s kind of like you and then hallucinates and has all these issues.”
So far, Eternos has signed 17 clients, with an additional seven who have expressed interest.
The company is privately funded and may raise some external financing eventually, but “we have no need for it right now,” LoCascio said.
This is one of a series of short profiles of AI startups, to help you get a handle on the landscape of artificial intelligence activity going on. For more on AI, see our new AI Atlas hub, which includes product reviews, news, tips and explainers.
Editors’ note: CNET used an AI engine to help create several dozen stories, which are labeled accordingly. The note you’re reading is attached to articles that deal substantively with the topic of AI but are created entirely by our expert editors and writers. For more, see our AI policy.