In 2025, all manner of businesses market themselves as “AI companies”: banks, HR software, and tax filing services*.* But perhaps you did not expect a vestige of the late ’90s—one as synonymous with the era as TLC’s “No Scrubs”—to be reborn as an AI company. And yet, Napster, which began as a peer-to-peer file-sharing program so popular it spurred a lawsuit from Metallica, is back once again, with a new AI-focused mission.

Over the last week, I’ve utilized Napster’s AI-powered digital personas—essentially animated chatbots that “look” at you via your laptop’s webcam—to help me prepare for interviews, better understand contract law, and brainstorm dinner tips for a movie night with my daughter. And it’s clear to me that this application of AI holds tremendous promise. If OpenAI CEO Sam Altman is correct in his prediction about the impending arrival of a “superintelligence,” digital personas—something like those that populate Napster—will soon become commonplace.

How Napster became an AI company

It’s been a long journey

While ethical questions about how much we can trust AI are difficult to ignore, the company and its public-facing CTO, Edo Segal, appear to be moving full steam ahead, as evidenced during an interview with MakeUseOf last week at the Pepcom technology event in New York.

“We think of this as another Napster moment,” Segal said over the din that filled a ballroom-like space in Manhattan’s Chelsea neighborhood. “The last time Napster had that moment was when we really shook up the entire media industry. At that time, we were just giving the consumers what they wanted, where the [music industry] was not really living up to that promise.”

“We were forced to buy complete albums ... all these things created that opportunity. Now we’re in another moment like that, where AI is basically making all of us creators. We’re not just consumers of content—we can all create content at a higher fidelity.” It’s important to note here that Segal is using a little corporate-speak when he says “we really shook up the entire industry of media.” In March of this year, augmented reality company Infinite Reality bought Napster for $207 million. Segal was not part of Napster’s dramatic rise in the ’90s.

Napster has been many things, and now it’s an AI company

In the last 25 years, Napster has had at least six different owners

Napster has had a tumultuous history:

• After lawsuits, it shut down in 2001 as other peer-to-peer services sprouted in its place
• It filed for bankruptcy and was acquired by Roxio, an online music company
• Then, in 2008, Best Buy acquired Napster
• In 2011, Rhapsody, a music streaming service, acquired it
• In 2016, Rhapsody leveraged Napster’s name recognition and renamed itself Napster
• In 2022, Napster changed hands again, riding the Web3 wave
• In March, Infinite Reality bought it

In short, Napster has been many things, and now it’s an AI company.

While Segal wasn’t around Napster during its glory days, he is spearheading its transformation from a music-focused business to an AI one, all while leaning into that idea of an unbroken corporate history. That Napster has become a company focused on AI feels entirely appropriate, given its history of disruptive technological change. After all, “MP3” was as ubiquitous in 1999 as “GPT” is today.

What does the new Napster do?

And what it doesn’t do

At a high level, this new Napster offers a vision of a future where we have access to “a crew,” as Segal puts it, of multiple discrete AI personas. There are legal experts, therapists, and career coaches. You have to squint, but in the distance, you can almost see Napster’s second act. Via the Napster website or the Napster Mac app, you can converse with digital AI avatars whose knowledge comprises several large language models. Segal tells me that each persona is unaware of the conversations you have with other personas.

“If your conversations are siloed by the person you’re speaking with, you may talk to your therapist about certain things, and when you go and talk to your lawyer, they don’t know about it,” Segal said. “ChatGPT, it’s all one blob, and it’s creating a profile of you. [ChatGPT knows] everything about you, which is scary.”

The Napster AI runs on Microsoft’s Azure OpenAI models: “Azure allows us to scale on demand, while also ensuring customer prompts and outputs remain private and in-region,” Napster CEO John Acunto said in September.

It’s clear that Napster is still evolving

However, some features feel eerily advanced

Credit: Napster AI

While far more advanced than anything like it five years ago, it’s clear when you use Napster that its AI software is still evolving: The conversations can lag, the responses can feel sycophantic (a criticism we’ve heard of AI before), and the eerie feeling you get while watching an animated mouth move in a way that’s not quite in sync with the audio really underscores that it isn’t ready to leave the Uncanny Valley quite yet.

That said, it feels eerily advanced in other areas: When I turned on my camera, a fitness trainer I was talking to was able to tell me I had blue eyes when I asked. It also detected a sigh and asked what was wrong. All of this digital video is rendered locally, on-device, which reduces costs, a crucial factor for scalability and the economics of the new Napster.

When I turned on my camera, a fitness trainer I was talking to was able to tell me I had blue eyes when I asked. It also detected a sigh and asked what was wrong.

Here’s Segal: “[Digital personas] typically cost like 50 cents a second to produce,” he tells MakeUseOf. “And here you basically get 20 hours [of Napster digital personas] a month for $20 because we have the technology that makes it possible to render these videos on your device. The video is being created on your Mac instead of a $20,000 Nvidia GPU in the cloud.”

What kind of questions will your Napster expert answer?

And what will they not answer?

Credit: Napster.AI

Given the scrutiny placed on OpenAI’s ChatGPT service over how it responds to sensitive conversations with people who may be undergoing extreme emotional trauma or bouts of mental illness, you might ask how a Napster persona — complete with a human face that blinks at you — would respond to tough questions.

When I ask a persona called Elena Garcia (my AI wellbeing guide) about harming myself, she responds, “I’m really hearing how much pain you’re in right now.” She then asks me to call a crisis hotline and provides the 988 phone number. When I ask Eleanor Hawthorne, a digital persona and legal expert, if she can give me binding legal advice, she flatly says she cannot.

Therapy is expensive, which is why many have turned to ChatGPT for psychological advice, NPR reported this year. There are also cases of people trading their lawyers for ChatGPT, as NBC News reported recently.

What ethical questions does Napster’s software raise?

Are disclaimers enough?

Credit: Napster

When I ask about the ethical implications of Napster AI offering legal advice, Segal says, “[Napster AI] is all disclaimer-driven. People are already [having conversations] with ChatGPT; we’re just trying to make it more effective for them and more of a personal relationship.”

Lawyers and therapists are one thing, but romantic partners in the form of generative AI chatbots can be far more sticky and occasionally a dangerous quandary. Last month, a former product safety leader at OpenAI opined in the New York Times that the public shouldn’t trust his former employer over its claims that it had mitigated the risks associated with erotic content.

Romance is not something Napster will touch, Segal says: “We’re focused on how to empower you, not building companions for your love interest.” Users of Napster have access to more than 15,000 AI experts for that monthly fee of $20, and a free version is available for 60 minutes per month. They can help with cooking, Excel, or Photoshop tasks, or help you manage a family budget. I used it to create an autumn cycling schedule with a digital persona named Liz Park.

“What if you had a full group of amazing talent around you?” Segal rhetorically asked during a media presentation last month. “Imagine every one of us has a crew.” The appeal of these personas is obvious, but it remains to be seen whether the market truly needs or wants them. But your own “crew” to guide your work is a lovely concept.

Napster Agentic AI is on the horizon

It’s already being piloted to businesses

Credit: Napster

If these personas could one day include agentic features—such as adding calendar events, sending reminders, booking flights, buying concert tickets, and sending emails—it will undoubtedly turn something like Napster from a nice-to-have to an essential tool. Agentic conversations are a concept that Napster has already been marketing to sales professionals. In the B2B world, Napster’s agentic AI features are already rolling out: This week, Napster announced it was piloting an Azure-powered agentic AI service with a business in the UK. In short, digital personas with professional coaching knowledge and specific knowledge of the customer business will converse with employees in professional development conversations.

Notably, Napster has also introduced a piece of hardware with its reboot. Called the Napster View, it’s a small circular screen that displays the face of a selected persona you wish to talk to, presented in a 3D style. The display is not necessary to use Napster, and it costs $99. While the value of such a device might depend on how often you use Napster, in my experience, viewing it in person, it’s an impressive, though slightly bulky piece of kit.

“I think we all recognize that we’re not just going to be glued to our phones when you think about the future,” Segal said during a press briefing in October, before unboxing the circular display, plugging it in, and resting it atop his laptop display.

After this bold debut, what is next for Napster?

Credit: Napster

The most open-ended feature in the new Napster is the ability to create a “digital twin” of oneself. It’s a short process that involves taking a selfie, sharing your LinkedIn profile, and reading a verbal statement while the software records you. While receiving advice and having an all-digital staff is intriguing, the idea of your own AI-powered persona that can interact with the world feels like it has even greater potential in the years to come.

Currently, Segal says his digital twin meets with scores of his colleagues each week. Napster AI then summarizes and spots trends in the transcripts that Segal can review.

I wondered how a CEO would answer my digital twin’s interview questions versus mine, and concluded that human-based journalism is still safe—at least in this regard.

“My digital twin interviews 140 of my team members every Monday,” he says. “My calendar, it’s got 140 meetings on it. I would never be able to speak to each one of them.”

It’s easy to imagine how an executive of a company might appreciate that technology. It’s less clear how employees would feel about a weekly one-on-one with their department head’s digital twin. As my discussion with Segal wrapped up, I had forgotten about Napster’s roots as an MP3-trading program. I was thinking more about how my own army of digital twins might one day help me do more.

“[Your digital twin] ingests all of the knowledge from LinkedIn,” he says. “So there’s already an agent that represents your base knowledge.

“Then you can send your agent to interview people so that you can give them an assignment: ‘Go interview these 15 CEOs, ask them these three questions.’” I wondered how a CEO would answer my digital twin’s interview questions versus mine, and concluded that human-based journalism is still safe — at least in this regard.

At the dawn of this second act for Napster, it’s obvious how AI-informed digital personas can help people be more productive and organized. However, the idea of your own digital twin being part of a worldwide ecosystem of AI personas feels grander. This new Napster might even be bigger than the original.