AI Listening Devices: Handy Helpers or Legal Minefields?
14:47 Share

This is a CRM Meltdown. Hey, boss, our chatbots glitched, 300 orders vanished, and everything got escalated to our live agents. Wait times are over two days long. Call me. Bad CRM was then. This is ServiceNow. CRM for the AI era. Hey, TNB listeners. Before we get started, a heads up. We're going to be asking you a question at the top of each show for the next few weeks. Our goal here at Tech News Briefing is to keep you updated with the latest headlines and trends on all things tech. Now,

Now we want to know more about you, what you like about the show, and what more you'd like to be hearing from us. We already asked you about some corners of tech you might be interested in. Now we've got a few others in mind. Biotech? Data science? Robotics? Let us know what sparks your interest. If you're listening on Spotify, look for our poll under the episode description, or you can send us an email to tnb at wsj.com. Now on to the show.

Welcome to Tech News Briefing. It's Monday, May 5th. I'm Victoria Craig for The Wall Street Journal. Weeks into a legal showdown between the Federal Trade Commission and Facebook parent company Meta, what have we learned? Our reporter gives us an update from what he's seen and heard in the courtroom. Then, would you willingly allow a device to record your every word, every day? Our tech columnist did for three weeks. She'll share her insights from the experience.

But first, is there enough competition in social media? That's the big question a judge is looking to answer without a jury in an ongoing legal battle between the Federal Trade Commission and Meta. The FTC argues that Meta created a monopoly by buying up its competition, WhatsApp and Instagram, more than a decade ago. But Meta says social media has become a lot more competitive since then, and it faces increasing competition from the likes of TikTok and YouTube.

Dave Michaels, who covers antitrust litigation and corporate law enforcement for The Wall Street Journal, has been in the courtroom as the trial unfolds. Dave, why is the FTC arguing that YouTube and TikTok don't actually compete with Meta's social media platforms? Yeah, the FTC says that those services, as well as X, formerly Twitter, are basically entertainment platforms.

Services or their apps that connect you with your interests. The FTC says those services are not built on what's called a social graph, meaning they're not built on the relationships that you as a user have with friends and family. And the FTC says that network of relationships is.

is what is at the heart of the value proposition for Facebook and Instagram, and even WhatsApp, which is a messaging service. But a lot of people use it in a way that they stay in touch with groups of friends or groups of family members within the app. That's important because if the court agrees with the FTC that the social media market for friends and family

just includes the companies that the FTC says it does, then the FTC has a good chance of winning. But if the social media market is broader and TikTok is part of it, then it will be a lot harder for the FTC to be able to convince the court

that Meta has a monopoly today. And I thought one of the interesting parts was what the Instagram co-founder had to say about the success or not of his platform before Meta bought it all those years ago. Yeah, there was some really compelling testimony from the Instagram co-founder, Kevin Systrom. And he was saying that Instagram was growing super fast and

It was doing something similar to what Facebook did, but with photos at the center of their offering. And he was insistent that after Meta bought Instagram, Meta's executives, including Mark Zuckerberg, were really conflicted about the effect of Instagram's growth on Facebook's. That is that

There was concern that they were competing for users and that the marginal new user for Instagram was perhaps not using Facebook as much as a result. Systrom eventually was questioned by Meta's lawyers, and he conceded a point that is important for Meta.

When he said that today, these products, these social media products, they're all about entertaining users with video content. That's what TikTok does. That's what YouTube Shorts does. That's what Facebook does. That's what Instagram does. So even Systrom conceded that the market for user attention today is very different than it was in 2012.

You also mentioned in your story the judge presiding over this case. Can you just give us some context around some of the questions that he's been asking both sides in this case so far? Yeah, it's U.S. District Judge James Boasberg and

On the one hand, it's pretty apparent that he doesn't use a lot of social media, and he has asked some questions that get at the heart of the issue, such as whether the differences between social media today are really firm or if it's just a – as he questioned it, it's just a difference in degree. In other words, these projects share a lot of features.

And so perhaps they're not as different as the FTC maintains. So, Dave, what's next in this case? We think there's at least a week more of testimony from the FTC's witnesses, including the FTC's main expert and expert witnesses, Kemp.

can play a really big role in these antitrust cases, trying to explain to the court why the market is defined the way that the FTC says.

And the expert has to be very credible in making that presentation. But then Meta will get its chance and it will put its witnesses up on the stand. It will question those witnesses and try to drive home its point that this stuff about Instagram and WhatsApp being acquired is just ancient history. It's going to try to drive home that point that

Meta today faces just an onslaught of competition from TikTok and YouTube and other forms of social media and entertainment that were just not as prevalent in 2012 when this market was making that transition from desktop to mobile. That was WSJ reporter Dave Michaels. Coming up, if we want artificial intelligence to become all-knowing assistants in our daily lives, we need to know how to do it.

we gotta let them learn by listening. At least, that's the mantra our tech columnist followed recently. After the break, she'll explain how it's made her life easier and how it might cross some legal lines. This is a CRM Meltdown.

So it gets worse. Inventory swears the units arrived last week. They bounce me to Warehouse who claim inventory's full of it. So either inventory's got a vivid imagination or Warehouse lost everything. Our manufacturer says they'll look into it. But the only guy who'd know anything about this is Bob. But Bob's in Bora Bora. Sure hope Bob's enjoying his mango tangos because we're about to lose so many customers. Bad CRM was then. This is ServiceNow. CRM for the AI era.

Want AI to generate a to-do list for you or maybe remind you to send that email you keep forgetting? How about summarizing a long meeting you zoned out in? Well, 15 years ago, I would have told you there's an app for that. Now I'll say there's an AI listening device that's always on for that.

WSJ Tech columnist Joanna Stern strapped one to her wrist for three weeks. Joanna, first, why in the world would you do this? Back in January, I read about this device called the Bee Pioneer Bracelet, and it was $50. It was announced at CES, and I saw people talking about this device that just records everything. It is nonstop recording when it hears your voice or other voices.

And then it uses AI to summarize and analyze what you are saying. And I will say that within a few minutes or hours of wearing this thing, I was completely blown away. One, by how creepy it was. The fact that everything I had said had been transcribed and now lived in an app. And that wasn't only everything I'd said, but anyone I had talked to.

that transcription was there. And then two, how it was proactively listening to those conversations and adding things to

that I had said I would do or somebody had asked me to do to a to-do list. And so my memories and everything I need to remember has just been outsourced to this bracelet on my wrist. And that's the part that I'm so curious about, because if you think about this device hearing absolutely everything you do in a day, what actionable information did it create for

for you from all that it was hearing you say and do. Here's an example. Me and you were talking about this. And at the end of this conversation, I'll say, I'm going to send you that recording. I'll do it when I'm done with this next meeting. And it would take that and put it on my to-do list. It would say, send Victoria the recording of the podcast. It has that context. It's using AI, using large language models to make sense of all that's being said.

A question that a lot of our listeners will probably have is how exactly can it process all of that information? Because you're talking about 24 hours a day, 365 days a year. What does it do with all that information and how does it process it so quickly? The way Bee does it, and I did review some other devices, but I'll talk about how Bee does it, is the device is listening. It's

monitoring for audio. Once it hears that audio, it uses Bluetooth to take the microphone, what it's hearing on the bracelet, sending it to the phone, and then sending it to the cloud to do the processing of the transcription. And B is actually tossing out the audio. You can't listen back to your conversations. You can only read the transcriptions. And frankly, they're not great transcriptions. We use transcriptions, of course, a lot in our work, and we want them to be as close as possible so we can get the quote perfect.

This doesn't really need to get the quote perfect. It just needs to get like a general sense of what that conversation was about. And so then they take that transcription and they run it through large language models. And they're using that to summarize as you go. So yes, it's a lot of data, but it's also not as much because you don't have that big audio file and it's been turned into a text file, which is smaller and they're doing...

special stuff to make those files smaller and then send it back to your phone and your app so you can always access it. What protections are in place to make sure that the private conversations that you do have, if you don't think about, because you mentioned in your story, there is a way that you can turn it off, make it stop listening, or is lack of privacy just something we're going to have to get used to as these things become wider spread?

Look, this is a privacy nightmare, but Bee and Limitless, which is the other one I've talked to, they have the right ideas about privacy in mind if you're going to build something like this. Yes, both of them have buttons on them to turn off the mics. Limitless is this little pendant that works in a similar way, but actually also gives you the audio to listen back to. It has a little light on it that's always glowing when you've got the mic on.

B is in the process of changing the light setup, but right now it does also have a button. You press it, and actually when you press the button to turn off the microphone, it glows red, which doesn't make much sense since the universal light for recording is red, but they are planning to switch that soon in an update. Another thing you can do is you can delete any conversations or your account from the apps, and they say that completely clears this from their servers.

Limitless, they keep that audio. You can say in the app how long you want them to keep the audio so they can toss it after a number of days or they have some presets that you can do. On top of that, both of them say that this information is encrypted on their servers. Nobody or no employees can access it. But of course, there's always risk with all of this stuff. For the tech savvy first adopters out there, though, one of the important things to consider...

as this, if this becomes much more widespread, is that there are state-by-state restrictions on recording people's speech without their permission. So is this a minefield for individuals to navigate? Whose responsibility is this? Absolutely. The first questions I started getting when I would tell people I was recording, and I was really clear with most of my conversations, like this is recording, it's not recording,

recording the audio with the bee and I would explain all of that. But almost always the first question I would get from people is, is that legal? Can you even record me without my consent? And so talk to a few lawyers about this. And it really depends on what state you're in. So I was in New York and New Jersey, which are one party consent states where only one person, that's me in that case, has to agree to the recording.

But if I were in one of the about a dozen other states that require two-party consent, I would need that permission of the person I was talking to and possibly people within an earshot that might be getting picked up. And I asked the companies about this. They said they are going to start to get more aggressive about telling people this. But there is such usefulness and utility in this that I do think some will start to use it. And I do think also I heard from a lot of readers in response, especially those that are starting to experience memory loss, and they were pretty game to try it.

That was Wall Street Journal's senior personal tech columnist, Joanna Stern. And that's it for Tech News Briefing. Today's show was produced by Julie Chang with deputy editor Chris Sinsley. I'm Victoria Craig for The Wall Street Journal. We'll be back this afternoon with TNB Tech Minute. Thanks for listening.

This is a CRM meltdown. Hey, good news! Only 200 customers are on hold now. Bad news! It's because 100 canceled their orders. We need a new CRM. Bad CRM was then. This is service now. CRM for the AI era.