AI Natives Are Now Entering the Workforce

[00:00:00] Searching for a home? The Real Commute Tool from Realtor.com lets you find homes right where you need to be. Find a home 10 minutes from work, 15 from school, and 20 from grandmas. With over half a million new listings every month, you're just minutes from finding your perfect home on Realtor.com. Trust the number one site Real Estate Professionals Trust. Search now on Realtor.com. Based on average new for sale and rental listings July 2024 through June 2025. Number one trusted based on August 2025 proprietary survey among real estate professionals.

[00:00:33] Welcome to Tech News Briefing. It's Friday, May 29th. I'm Isabel Busquette for The Wall Street Journal. AI is often criticized for spreading misinformation, but it turns out large language models are actually surprisingly good at talking people out of conspiracy theories. We're diving into why and what we can learn from them. Then, here comes the class of AI. Hi. This year's college graduates are so-called AI natives, and they're now set to enter the workforce.

[00:01:03] We're breaking down how the job market is treating them and whether they're uniquely disrupted or advantaged. But first, new research has found that in some corners of the internet, AI is helping combat misinformation.

[00:01:21] So-called debunk bots or chatbots designed to fact-check conspiracy theories are successfully changing people's minds by generating facts and evidence quickly and presenting it in a clear, understandable way. WSJ contributor Lisa Ward joins us for more on how they do it and what we as humans can learn from them. What does the research show about AI's ability to talk people out of conspiracy theories?

[00:01:51] They're actually pretty good at changing people's minds. In one of the studies, participants discussed conspiracy theories they believed, and the researchers asked AI to persuade participants that those theories weren't true. Most participants not only changed their minds immediately after the conversation with the AI, but also continued to hold those new views for about two months afterwards. What types of conspiracy theories did it talk people out of in that study?

[00:02:17] All types of conspiracy theories, from 9-11 to Kennedy's assassination to conspiracies based on anti-Semitic tropes. In one study where the debunk bot talked people out of anti-Semitic conspiracy theories, participants actually saw Jewish people more favorably afterwards. So it wasn't based on a specific conspiracy theory.

[00:02:38] They tried all different ones, and I think they focused on anti-Semitism because they thought it was, you know, a deeply held belief, and they were curious if you could challenge those facts, would it make a difference? And they were surprised it did. So what is it that makes the AI so good at this?

[00:03:16] Does the tone that the AI is talking to the person in matter at all here? It only matters if the AI is condescending. You know, people really didn't like that. But there is no difference in effectiveness between a neutral or affirming debunk bot. You know, this is a great use case, but are there also concerns that AI could further perpetrate misinformation or conspiracy theories? Yeah, so in these studies, the debunk bots were instructed to really challenge conspiracy theories.

[00:03:46] But with commercial large language models, they're not always given those specific instructions. So it's unclear to what extent these models are going to correct you. So it's important to be cognizant of that possibility. So what can we as humans learn from this about the best way we could maybe talk each other out of conspiracy theories? I mean, being able to marshal a lot of facts and evidence to refute conspiracy theories is really important. Of course, that can be difficult.

[00:04:14] Often conspiracy theorists will say, what about this piece of evidence or that piece of evidence? And you need to be able to have an answer for each of those things. And that can be difficult to do. Also, some pretty basic manners matter, too. If you're going to get a person to engage, you shouldn't insult them, communicate that they're an idiot or start yelling. It just shuts the conversation down. That was WSJ contributor Lisa Ward. What are some other ways you've used large language models? If you're a listener on Spotify, leave us a comment with your answer.

[00:04:45] Coming up. Throughout their college careers, the class of 2026 has received a lot of mixed messaging about AI. How's it impacting them as they enter the workforce? That's after the break. It's graduation season.

[00:05:09] But the class of 2026 is entering one of the toughest job markets for young college graduates in years. Young adults are facing elevated unemployment rates as companies are increasingly using AI to handle the entry-level tasks that once served as career launch pads. Still, more than any other age group, young graduates say they feel prepared to compete in an AI-shaped job market.

[00:05:35] And many employers say this graduating class brings a unique set of skills that could actually give them an edge. WSJ reporter Allison Poehli joined our colleague Imani Moise to talk about what this moment means for the future of work. You spoke to several graduating seniors for this story. What's unique about the class of 2026?

[00:05:58] So they were in their first semester of freshman year when ChatGPT was released to the public. And then as they progressed through college, AI went from being something that they were told don't use, it's cheating, it's plagiarism, to something they were encouraged to use. And then were told this is either going to help you get a job or it's going to take away all the job opportunities that you thought you had.

[00:06:24] So what's interesting about the class of 2026 is they kind of, in some ways, came of age with AI and have developed along with it over the past few years. And now they're entering the job market at a time when it's all anybody is talking about. How would you sum up their feelings towards AI? There's an ambivalence. I think some of them recognize that there's this inevitability where the cat's out of the bag. It's going to be an expectation in the workplace.

[00:06:52] At the same time, there is this resistance to it where people are concerned about relying on it too much. They're worried about what it will do to their critical thinking and what it will even mean for their careers going forward. And how does that compare with the way employers are thinking about AI right now? A lot of employers are all in on AI. And people's performance is now graded on how much they use AI at work at various financial firms and other companies.

[00:07:21] The employers are expecting that these students will come in knowing how to use the technology, that they'll be AI native, so to speak. And so it's a skill that certain companies like Salesforce, for example, are looking for very specifically in the entry-level hires that they're bringing on. What companies are most eager to hire these AI native grads? Is it mostly happening in tech or are you seeing it spread across more industries?

[00:07:46] It's definitely happening in tech, but I talked to some students who were applying for jobs in graphic design and in HR. And their job postings, it said familiarity with AI tools as one of the requirements. So it is across the board, but there are notable companies, IBM, MetLife, Salesforce have all said we're really eager to hire people who are AI fluent. Consulting firms like McKinsey say they're increasing hiring as well.

[00:08:15] It sounds like a lot of these grads might be more fluent in AI tools than their managers. Did you get the sense that this could reshape traditional workplace hierarchies? I think there will be a reshuffling of roles and what we traditionally think of as certain jobs. But I don't know about the hierarchy because at the end of the day, you still need somebody who's going to be making a decision and making that judgment call.

[00:08:40] And just because someone can use a tool or is more fluent in helping someone get familiar with the technology doesn't mean they're in a position where they can be making the final call on a client-facing project or anything like that. There's still going to be managers who you have to report to. It kind of reminds me of being a teenager in the 2010s and social media is becoming a phenomenon.

[00:09:07] And honestly, my parents didn't know what I was doing on the computer. I could have been doing anything. Yep. Are you expecting some of that to happen in the workplace as well? Yeah. And I think that was something that came up in one of the interviews we did with the CEO of Shark Ninja where he was basically saying we need to check in with these entry-level hires more frequently because we can't just say,

[00:09:30] oh, go out and use these AI tools and come up with something and leave them alone for one to two weeks because they can get so far down a rabbit hole and be going in the wrong direction that it's like, oh, no. We just spent two weeks working on something that was totally irrelevant. Five or 10 years from now, do you think we'll look back on the class of 2026 as uniquely disrupted or uniquely advantaged?

[00:09:51] You know, when I was looking at the recent graduate unemployment data compared to 2020 and 2008, it's not as high as those two time periods, even though entry-level grads are struggling to find jobs. A lot of employers are saying, you know, we're not going to slow down on entry-level hiring. We still need these graduates.

[00:10:16] But in five to 10 years, entry-level hiring will look super different. I think this will be a class that was advantaged in that they had a decently good market even though it was still challenging. But things are going to continue to be disrupted in ways that I don't even think are fully realized yet and we can't predict. So maybe both. That was WSJ reporter Allison Pohley speaking with our colleague Imani Moise.

[00:10:44] 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, Isabel Busquette. Jessica Fenton and Michael LaValle wrote our theme music. Our supervising producer is Katie Ferguson. Our development producer is Aisha Al-Muslim. And Chris Dinsley is the deputy editor of audio for The Wall Street Journal. We'll be back later this morning with the TNB Tech Minute. Thanks for listening.