Can AI Find Your Next Summer Romance?

[00:00:03] Welcome to Tech News Briefing. It's Friday, August 7th. I'm Katie Dayton for The Wall Street Journal. Anybody who's spent an evening swiping through dating apps has at some point thought, there must be another way. And now there is. We're looking at why some daters are turning to AI matchmakers and how successful they've been in finding love so far. Then, we're diving into a Wall Street Journal investigation that explores how ICE is weaponising social media against its critics.

[00:00:34] The agency has deployed a round-the-clock digital surveillance programme to scour platforms including X and Reddit for threats to its agents, looking to uncover the identities of posters. We're unpacking how the initiative is working and how social media companies are responding. But first, many of us are entrusting AI with our calendars, our meeting notes and our daily internet queries. But what about our love lives?

[00:01:01] For some frustrated daters, a new batch of AI-powered dating apps is allowing them to narrow down their choices and even cut out those awkward getting-to-know-you chats before first dates. My colleague, Isabelle Busquette, a reporter for the Wall Street Journal Leadership Institute, spoke to WSJ's Molly Reinman, who's been researching the trend as singles across America look for a summer romance. Molly, what are AI dating apps and how do they work?

[00:01:29] Yeah, so we've seen a few iterations of this since very user-friendly AI came on the scene a few years ago. Like AI clone-based dating apps where you would create a chatbot that was meant to talk like a dater and then those chatbots would go on these kind of virtual dates and those would be assessed for compatibility. Those kind of shuttered very quickly because the short of it is that the generative AI was just not really good at fully encapsulating a human themselves.

[00:01:58] So the new wave of apps that are on the scene now and that are really making headway and sort of inspiring some changes in some of the legacy apps, I've kind of been calling them AI matchmaker apps. And so how they work is you download the app, you talk to a chatbot for anywhere from 15 minutes to an hour, but you basically are just telling it everything about you, things that maybe a traditional dating app wouldn't ask and wouldn't know. Like what are some of your favorite restaurants? Why did your last relationship end?

[00:02:26] And it's meant to feel like a human matchmaker. So someone that's trying to get to know you in order to pitch you the best match. The AI matchmaker will then find you a match that you're compatible with and that's done using AI enhanced but still algorithmic matching. Then when it identifies someone that it predicts you would be compatible with, it'll come back to you with an AI summary of that person. And then from there, if both parties say they're interested, you pay per date.

[00:02:53] So you pay usually like $15 to $25 in the app, books the date, tells you where to go, you show up. So from the people you talked to, what did they like about these AI dating apps and what were some of their complaints? People liked that it was a lot less work than regular dating apps. The Forbes study from last year found that people on dating apps spend an average of like 51 minutes per day swiping. And that's obviously time suck.

[00:03:20] It's also quite frustrating, I think, when you spend that time and it doesn't go anywhere. And so these apps, you just do your onboarding call, you leave them, and then they'll reach out to you and they have someone ready for you. So I think users really liked that element of it. They didn't really have to put in a lot of work to get the payoff. And they also just liked that it felt like it got to know them better and things that were more relevant to actually finding them a good partner.

[00:03:43] I think a lot of the user issues were related to both it being very new and the tech maybe not being fully there yet to be able to like really make completely, consistently accurate, informed matchmaking decisions. So like users had experiences where they were like getting pitched people and they were like, I'm not attracted to this person. Or as we've seen with other sort of AI, it has a tendency to just affirm users. And so they felt like sometimes the summaries were just saying things they wanted to hear.

[00:04:12] The other biggest complaint people had, which is the opposite side of the coin to having to spend less time on the app, is they just weren't getting matches as frequently. So, you know, on Hinge, you can swipe through if you have a premium subscription, like infinite profiles a day. And these people were getting like one or two a month. Any early success stories you found? Any couples that are still together having matched on these apps? Yeah, I talked to two, which is really sweet.

[00:04:39] A couple who met on an app called Amato, which is live in New York. And another who met on an app called Known, which is live in San Francisco. And both of them said that the ability of the AI matchmaker to give them a summary of the person they were going to meet, but not allow them to talk to that person before, made their first dates feel a lot more natural and low stakes,

[00:05:04] which I found interesting because you might think that talking to a person themselves would do that. But no, they felt like they knew the core of the person because of their AI matchmaker's description of them. So what does it all mean for singles today? Is AI helping create more matches or do you get the sense they are feeling more doomed than ever? Gosh, it's hard to say. I think there's not a critical mass of people on these AI matchmaker apps to really be able to answer that question.

[00:05:32] But I do think that I've spoken to so many singles who are just fundamentally burnt out from the way these swipe apps work. So I think anything new is exciting to singles. A lot of people that have made the choice to try these apps made that choice for a reason. And that's because they wanted anything other than what they currently have. That was Molly Reinman, who reported on this for The Wall Street Journal, speaking with Isabel Busquets,

[00:06:00] a reporter for The Wall Street Journal Leadership Institute. Do you think AI is improving dating apps? What has your experience been with them? If you're listening on Spotify, leave us a comment. Coming up, a new ICE operation is going down, not on the streets, but on social media. That's after the break.

[00:06:28] Immigration and Customs Enforcement has deployed a round-the-clock digital dragnet designed to protect its agents from threats and stay apprised of activities that could disrupt ISIS operations. That's according to a new investigation from The Wall Street Journal. The initiative has put pressure on social media companies as the federal agency asked them to turn over the identities of critical posters. WSJ reporter Shane Shifflett is here to explain what's happening behind the scenes.

[00:06:56] Shane, why exactly does ICE say that it's taking this action? ICE says that its agents have faced escalating threats since they've moved into cities across the United States to enforce immigration policies under the Trump administration. And as part of that, they say that their agents and the agents' families are having their locations revealed, where they live, their phone numbers put online,

[00:07:24] and they also say that online users are making threats that suggest individuals should take violence against immigration agents. And from a legal standpoint, can you take us through where the First Amendment sits here and how that's being used in defence? A lot of the things that happen on the legal side of these investigations involve filing subpoenas with the social media companies themselves.

[00:07:49] The users who made the posts, they may or may not know that there's been a subpoena filed with the social media company asking the company to turn over their personal information, things that would identify them like a home address and their name or an email address, for instance. When users are notified by the companies, some of them have obtained lawyers and the lawyers have pushed back. And in most of those cases, if not all those cases, ICE will drop its case.

[00:08:15] There are also grand jury subpoenas where a jury actually hears the agency's arguments or investigators' arguments and might compel a company to turn over some of these records. And they have in some cases. And to be sure, the Supreme Court has also held that true threats aren't protected by the Constitution. So if somebody is online making a real threat that's directed at a specific agent that's credible, then that's not protected speech.

[00:08:42] In the case where ICE has succeeded in unmasking its critics, what have its agents' next steps tended to look like? There have been a couple of cases where individuals have made true incredible threats. An Oklahoma man was convicted of threatening to kill federal agents after he wrote, we as U.S. citizens should be gunning down these domestic terrorists on X, along with several other posts about the agents needing to be shot. There were also two women who were found guilty of stalking in Los Angeles

[00:09:11] after they followed an ICE agent's vehicle while live streaming on a popular Instagram account and ended up at his home. Both of those individuals said that they thought his car was headed to an immigration raid. But this, you know, is one of the categories where ICE is concerned about revealing personal information about its agents. But the Justice Department declined to elaborate on any of these cases. You know, there's other cases where the line isn't so clear, where ICE has implied or said that there's a threatening message online,

[00:09:40] including the case of one person in January who wrote on X that all ICE agents need to be put down like the dogs they are. ICE agents looked into this, investigators looked into this, and they ended up showing up at her door. They determined her address after a court ordered X to turn over her information. So these plainclothes agents, they showed up and they told her not to post like that again. And then they left. And she hasn't faced charges, but she has curtailed her speech online.

[00:10:07] How have the social media platforms been handling these requests from ICE? Social media policies on how to handle these administrative subpoenas and requests for information differ depending on who is actually requesting the information. Sites like Reddit, they will scrutinize these requests. And if they determine that the users were engaged in an activity protected under the First Amendment, they will object to these requests. Reddit told us that privacy is central to how Reddit operates.

[00:10:38] And the company routinely objects to the government's requests, especially when it threatens a user's privacy or civil rights, including the First Amendment rights. Taking a step back, have we seen any government agency act in this manner before? Or is this a pretty unprecedented plan of action for ICE? DHS has monitored for threats online against its agents for years now.

[00:11:02] Now, this, according to the sources that we spoke with, which include current and former ICE agents and contractors who have worked in sort of open source surveillance areas, have told us this is definitely an escalation. And this involves much more parsing of the intent and meaning of the speech that users are making, much of it seemingly innocuous.

[00:11:26] Yet agents are showing up at people's houses and they're increasing their legal activity to try and obtain users' information. That was WSJ reporter Shane Shifflett. And that's it for Tech News Briefing. If you're a listener on Spotify, be sure to leave us a comment. Today's show was produced by Julie Chang. I'm your host, Katie Dayton. Jessica Fenton and Michael LaValle wrote our theme music. Our supervising producer is Katie Ferguson.

[00:11:55] Our development producer is Aisha Al-Muslim. And Chris Dinsley is the deputy editor of audio for The Wall Street Journal. Tune in tomorrow morning for the second episode of our special series, AI and the Blurring of Reality. Thanks for listening.