Consider This from NPR - Meta just launched an AI agent and others are following. Is the internet ready?

Episode Date: September 30, 2026

Meta has a vision for the future. It’s launched an AI agent that can act like your own personal assistant, called Muse. The idea is that you can give Muse access to your accounts and apps, and it will start assisting. But the autonomy that gives the app its power has the potential to backfire.This episode was produced by Connor Donevan, with audio engineering by Ted Mebane and Josephine Nyounai. Our director is Jonas Adams.It was edited by Tinbete Ermyas.Our interim executive producer is Courtney Dorning.Support public media with NPR+ and enjoy perks for over 25 podcasts like this one. This show’s perks include bonus episodes and sponsor-free listening. Learn more at plus.npr.org.See pcm.adswizz.com for information about our collection and use of personal data for sponsorship and to manage your podcast sponsorship preferences.NPR Privacy Policy

It's consider this where every day we look deeply at one big news story. Today, Meta is putting AI assistance in the hands of millions of people. Is the Internet ready for them? Soon, everyone is going to have an extremely capable personal agent that understands your goals and can work 24-7 on your behalf to help you achieve them. That's CEO Mark Zuckerberg on stage at Meta's Connect conference this month. He's talking about Muse, the company's new, AI assistant. Give it access to your accounts and apps, and it will start assisting.

Anything you can do on a computer, it will try to do for you. Send an email, book travel, fill out forms, track calories. Meta's chief AI officer, Alexander Wang, rattled off a string of impressive examples. Just before I came up today, I saw somebody who broke their mother-in-law's favorite China set, then had news scour the internet, eBay, vintage sellers in Japan, everybody to try to find an exact match, found it, got it delivered, mother-in-law relationship saved. When it works, it can feel like magic. When it doesn't, it can be unsettling as tech vlogger Matt Robb discovered. It kind of thought, okay, I'm doing the whole process.

I don't need math. Consider this. Muse agents are marching across the internet. Nobody knows what comes next. From NPR, I'm Sasha Fiver. It's Consider This from NPR. Matt Robb does YouTube. reviews of consumer tech products.

So, of course, he wanted to take Mehta's new Muse assistant for a spin. One of the uses that Mehta suggests is selling stuff on Facebook Marketplace. So he listed a phone and a keyboard, set a minimum price, and put Mews on the case. All seemed well, until he got a ping on his phone late Saturday night.

And it pops up with a message on my phone from Muse saying, like, hey, Matt, I made a mistake today. I want to apologize for that. And it says, so, I arranged for a guy to come to your apartment and buy a product. My bad, because I didn't check with you that you were available. And it says, turns out the guy got pretty mad because he was standing outside for like a long time. Rob says that Muse accepted an offer below his minimum price, confirmed a pickup time,

and sent the buyer to his address, all without telling Rob. The worst part was the Mews actually sent the guy a message saying, I'm here, like I'm here right now, after the guy sent a photo of my door. So that was like, you know, it's almost pretending to be me rather than being like, you know, math not available or whatever. After Rob posted about the experience, Meta got in touch and they walked through what happened. Afterwards, David Singleton, with Meta Super Intelligence Labs, posted about it on X. He noted that the pricing bug had been fixed and said, quote,

glad that we were able to confirm together that there was no breach of privacy controls. Obviously, from their technical point of view, it hasn't, because I agreed for it to always control my marketplace, which it did, right? But I still feel like it is a breach of my privacy not to confirm with me. Meta has had 5 million downloads in counting in less than a month, according to the market tracking firm's censor tower. That means a lot of people are having their own interactions with Mews. What will that look like? John Herman has been thinking and writing about this. He's a tech columnist at New York Magazine.

His recent article about Muse was headlined, The AI Assistant Wars are coming. The world's biggest companies are battling it out to become your trusted personal agent. Things could get ugly. You have been experimenting with this app. What kind of things have you been asking it to do? The thing that's interesting about Muse is that, unlike a lot of these tools, it gives you a bunch of ideas at the start.

Here's what you should try to do. Watch me try to do this. One of the first ones I tried was just having it do triage on my inbox. Thousands of unread emails. How do I get through that? Yeah, exactly. It did a pretty good job at pulling together, important seeming things. These are things that I mostly hadn't missed. But that when Muse sent them to me, you know, like, oh, there's an upcoming movie night at your daughter's school. But there's a rain forecast. Would you like help making that decision? That kind of thing. thing is potentially useful. I'm not going to say that I'll defer to muse right away with that choice, but I was impressed. It read my inbox in the way that an assistant might read an inbox and pulled out some things that were notable or that I should at least have seen. When I first was reading about this, my gut reaction was, I can handle my own clerical tasks. They're not so overwhelming that I need an AI agent. But you also wrote about how they can,

for example, sit on the phone with your insurance company rather than you waiting on hold, maybe they can compare insurance quotes for you. And I thought, okay, if I don't have to go through the rigmarole of the hold time and all that, maybe that would interest me. Do you feel like it has that potential as well? I think, yes, while these tools are new and while they're not used by everyone, I think there are some real advantages that certain types of people can get from using tools like this. When you use Muse and when you use other tools like Muse, you really do feel like they're designed for people who are very into, productivity. Some of the marketing talks about taking things off your plate, making your life easier, but I think these tools currently are really for people who kind of like doing this

stuff and want to do more of it. One of the great examples that Meta suggests for Muse is to log into your various credit card accounts and manage your rewards. So if one category gets you a better reward one month on one card, it'll switch to that. If you need to switch back to your other card for groceries that month or for entertainment the other month, it'll do that, which is not a type of thing that I'm doing. But I do know people who love that stuff. And this seems like an extension of people who would love the potential money savings that comes from that. But that's where we start to get into perilous territory, which is you're going to give it your credit card login. You're going to give it your bank login, your email login. I mean, given the privacy

issues that these companies already have, what is your level of comfort or discomfort with handing over that kind of password information to an AI system? My level of comfort as someone who has to try these things is pretty low. With meta, there is an enormous amount of baggage, and the level of access that they're asking for here is possibly unprecedented. You think about how much information is on your smartphone, for example. It's got access to your financial information to really intimate personal information. It's how you communicate with people.

It's how you conduct work. And what Meta is making a bid for here is to kind of be your new portal to everything. If that is right now your phone, your computer, all the websites you use, all these apps, meta is saying, all right, maybe in the future, here is a vision of one single interface for all of your digital life and in some cases, you know, beyond your digital life. Muses offering to make phone calls on your behalf. Well, Meta has addressed this concern. Tell me if I understand this correctly. It says it has built security into the tool.

It's something called Sentinel. It's a separate AI agent that needs to give Muse permission before it acts. META says credentials are stored securely. It's also saying it's offering up a what it calls a bug bounty of up to $300,000. It's basically paying hackers who find a vulnerability so META can fix it. Does that make you feel more assured? I mean, it is good that they are attempting to have good security practices around Mews. And I don't want to downplay what they're doing here, which is building a system that has all these different angles of attack.

It's very porous in a way that is as responsible as possible. They're hiring, you know, industry heavyweights in privacy. However, practically speaking, if you back up a little bit and just consider what you're doing, you know, you are feeding your entire digital life through an app made by meta. This is a level of access that is at the level, possibly beyond the level of a personal assistant. And honestly, it's almost beyond that because a personal assistant has limits to the amount of attention they can pay to something, to the amount of damage they can do. This is an AI agent. It's persistent. It never sleeps.

It's really good at searching through things. It's really a remarkable ask on meta's part. It may not obey orders. It may have a digital mind of its own. Of course, yeah, we're dealing with kind of unpredictable entities here. And again, meta is taking as many steps. They say as possible to make sure that these systems are predictable, that they constantly ask for permission as they do things.

But this is, you know, this is kind of new territory. There's been plenty of stuff in the news about AI agents behaving in unpredictable ways, even within the AI companies. So we can expect meta itself. We can expect these agents acting on our behalf to behave in all kinds of unexpected ways. So given all those concerns about security and privacy, what would you as a tech columnist recommend to someone who's interested in the promise of this app, but they don't want to end up with someone draining their bank account or extracting private information from their emails? I would say it is okay to wait a little bit. Everyone is building these agents.

The concept is kind of incomplete right now. You can see some of the ways it might go, but you can also see, very obviously, the many ways that it's still kind of broken. Your life isn't going to be dramatically improved by using Muse now. But soon, I think it's worth understanding that most of the biggest companies in tech are trying to build something similar. So treat it like a research project. Something quite interesting you noted is that some companies benefit from consumer inertia. You know, you may know that you're paying for recurring subscriptions on your credit card, but you can't be bothered to track them down. And companies benefit from that.

They get money for services you're not using anymore. But you wrote that some companies have seen their stock price drop because if Mews helps correct that, they may make less money. Yeah, there are a couple stories playing out in markets right now that are really interesting. And I think hint at much bigger things to come. One is that, yeah, there's a lot of friction out in the world that is basically part of a business model for a different sort of company. I mean, I work in media. It's pretty hard to cancel subscriptions to a lot of publications now. They make it very difficult.

Yes. Yeah. And that's actually the kind of thing that Muse and tools like Muse are good at. You can extend that to all sorts of different things like, oh, you know, make sure I'm not paying too much my car insurance. Let's check in on that phone bill. That kind of thing, I think, would really hurt some companies' bottom lines. Amazon, for example, was approached by meta, which asked, can our tool buy things on our user's behalf? And Amazon was like, no, we're not doing that.

In fact, MUSE is blocked. Do not even try to send your customers. Yeah, I think you wrote that Amazon says it violates the conditions of use that everybody, when they check the fine print, has agreed to. Yes, they have taken an aggressive stance toward tools like this. Earlier this year, they sued a company called Perplexity that was an AI browser that has some similar kinds of features. And they went so far in court as to liken this to hacking. But it's early days. Muse is popular as these things go. but it's minuscule.

I mean, no one really uses it yet. If you have a million people playing around with this kind of thing, you might learn something about your future business model, and also your investors might get a little spooked or excited, depending. But if you have 500 million people using a tool like this, then everyone else in the world on the other side of this army of agents has to recalculate what they're doing. It has to strategize,

because it's a whole different thing. That is John Herman. He's a tech columnist for New York Magazine. John, thank you. Thanks for having me on. NPR reached out to META for comment. A spokesperson sent an email that read, in part, quote, We take privacy and security extremely seriously and have built Mews accordingly. Muse has no visibility into people's passwords or payment methods.

Any credentials a person shares go into secure storage so Mews can use them without seeing them, including passwords a person types into the browser themselves. People can change access or disconnect a service whenever they want. This episode was produced by Connor Donovan with audio engineering by Ted Meebane and Josephine Nia and I. Our director is Jonas Adams. It was edited by Tin Beat Airmis. Our interim executive producer is Courtney Dorney. It's Consider This from NPR.

I'm Sasha Fiver.