LEILA FADEL, HOST:

The AI enhancements LinkedIn promised its users aren’t going as planned. The company has quietly rolled back a series of prompts premium users were seeing on the site. These prompt questions would frequently appear throughout a person’s feed alongside the gold sparkle icon to signal it was AI. Christopher Zara is a senior editor with Fast Company, and he’s been following these developments. Good morning.

CHRISTOPHER ZARA: Hi.

FADEL: Why did LinkedIn remove these prompt questions that were showing up in people’s feeds?

ZARA: So I asked them that, and they essentially told me that, you know, this was seen by them as a test of different AI enhancements that they’ve been rolling out, mostly for premium members. They didn’t say specifically why they removed them. Now, I asked them if this was in response to any criticism or underuse of the feature, and they told me flat out, we didn’t remove these because of any criticism. But the criticism was real, and these questions did disappear within a matter of months. So, you know, I think we can assume because they removed them that the test wasn’t going so great.

FADEL: (Laughter) And people had been complaining, at least publicly, about it, right?

ZARA: Yeah, for sure. I mean, this - they were being complained about quite a bit by the premium members. These are members who typically pay, you know, a monthly fee to have different features. And the one thing that was interesting is that these AI features, you couldn’t disable them. So people had - were making the observation that they were essentially paying for these features that they didn’t want. And, you know, if LinkedIn is going to put these in there, at least give premium members who were paying the option to disable them.

FADEL: So what are some examples of AI prompts that were bothering their LinkedIn premium users?

ZARA: Sometimes, like, for example, if it was a post by us, it might say, you know, what’s the origin story of Fast Company? It would be, like, the prompt question that would appear underneath. You know, I had posted a story about a recall. There was a big recall of candy earlier this year, and one of the prompt questions was, how significant is the impact of product recalls? I actually clicked that prompt, and it just basically resummarized my article. And...

FADEL: You’re like, I already know this. I just posted it (laughter).

ZARA: Yeah, yeah - and barely reworded it. Like, you know, why am I clicking this for more information that I already had given to LinkedIn? So with stuff like that. And some of the questions were pretty comical. I mean, there was one about pumpkin spice (laughter), where it was just like, you know, what can you gain from pumpkin spice? And...

FADEL: Oh.

ZARA: It barely makes sense.

FADEL: What can you gain from pumpkin spice? Fall joy (laughter).

ZARA: Yeah.

FADEL: Well, but LinkedIn clearly does have ambitions to use AI. What is happening with AI and LinkedIn right now?

ZARA: Yeah, I mean, they’re doing a lot. They still have AI prompts in the job listings for premium members. And you can click certain prompts and ask, you know, whether you’re a good fit for the job. It might summarize some of the qualifications. So I think there’s a lot of potential use for AI and more broadly for generative AI features. I think what’s happening in some cases is that as companies like LinkedIn or its parent company, Microsoft, or Meta or Google - you know, as they rush to release these features, what they’re doing in some cases is disrupting the core experience that users expect.

FADEL: Christopher Zara is a senior editor with Fast Company. Thank you for your reporting and your time.

ZARA: Yeah, thank you for having me.

(SOUNDBITE OF FAR ORANGE’S “WHEN WE MOVE”) Transcript provided by NPR, Copyright NPR.