AI Could be Key to Facebook Parent Meta�s Turnaround

Meta, the parent company of Facebook and Instagram, had one of its toughest years in 2022. It�s facing fierce competition and a hit to its advertising business. But artificial intelligence that processes the company�s data to better pinpoint user interest could be about to turn that around. Wall Street Journal social-media reporter Salvador Rodriguez joins Tech News Briefing host Zoe Thomas to explain.

This transcript was prepared by a transcription service. This version may not be in its final form and may be updated.

Zoe Thomas: It’s been a rough year for Meta, the parent company of Facebook, Instagram, and WhatsApp. There’s been increased competition. It faced blowback from users when Instagram made changes to its format. And Apple’s privacy changes have hit its ad business just to name a few. But the tide may be turning for the company as it uses artificial intelligence to crunch data it has and improve some of its struggling features. That’s according to internal documents reviewed by the Wall Street Journal and interviews with people familiar with the matter. I’m Zoe Thomas from the Wall Street Journal and with me to discuss how Meta is using AI and where it’s making a difference in its operations is WSJ Social Media Reporter, Salvador Rodriguez. Hi Sal. Thanks for joining me.

Salvador Rodriguez: Thanks so much for having me.

Zoe Thomas: Can you tell us what areas Meta has turned AI for?

Salvador Rodriguez: Well, so Meta uses AI all across the company, but in the last year they made it a priority in two specific parts, and that’s in improving the Reels that are shown to users in their Instagram and Facebook feeds, as well as using AI to better draw predictions for what type of ads to show to specific users. So finding deeper correlations from the data that they have from their own systems than they could previously.

Zoe Thomas: Okay, so let’s break those two things out, starting with Reels. How exactly is AI helping Meta here?

Salvador Rodriguez: In the last couple of years they’ve seen the upstart TikTok rise up and whenever you see a TikTok, their AI, their machine learning technology, it’s doing a really good job getting a sense of what you like and then showing you more of that. And that’s really good for optimizing and maximizing how much time a user spends on that app. And so Meta really had to level up and get to a point where they can rival TikTok in the sense of creating that type of flywheel where they’re showing you the right kind of reels and they’re learning from you which kind of reels to keep showing you moving forward.

Zoe Thomas: The other area that you mentioned is advertising. That’s been a big problem for Meta and for other social media platforms since Apple’s privacy changes. How is Meta using AI in this area?

Salvador Rodriguez: When Apple made these privacy changes, it really reduced how much Meta could rely on data from outside of their apps and services to determine which ads to show to which people and how effective those ads were. One technique that the company has turned to make up for that loss of revenue and that loss of data signal is they’re using more AI to draw deeper correlations from the data that they already have. So basically using more technology to better understand their users without actually collecting more data from them. So this is something that the company has really bet on. That’s what my sources have told me, and it’s something that they’ve been working on over the past year. They’ve seen progress and they hope to get to a point where they can make up for most of that loss of signal that occurred because of Apple. We know this from internal documents posted within the company, hearing about town halls at the company and our own sources who have been at the company, who have been part of this work, familiar with this work, and just a combination of all that, letting us know what the company’s strategy is for overcoming these two big challenges.

Zoe Thomas: Just to be clear, the AI that Meta is using, it’s not the generative AI like ChatGPT uses and that we’ve been hearing a lot about lately, right?

Salvador Rodriguez: That’s right. It’s very different than that. AI comes in many shapes and forms, and in this case it’s basically Meta having to improve their own systems that just get a better sense of what their users are doing and how to better serve those users. So it is AI, but it’s not the same type of AI as is en vogue at the moment.

Zoe Thomas: This other area that they’ve been struggling with has been the metaverse. We’ve talked about it before on the show. They made this big pivot. CEO, Mark Zuckerberg said they were going to be a metaverse first company, but where do things stand with that effort?

Salvador Rodriguez: Well, the last that we checked in regards to the metaverse, the company was not doing great. They’ve made Horizon the key part of their metaverse push, and from some of the reporting that we did late in 2022, we saw that they’ve actually lost users since launching, and that’s never good, especially for a service that is so new. I’m eager to find out how it’s done since then. And of course, whether there was any boost from the holiday sales period when typically more people go out and buy new Quest headsets. So we’ll have to see if there’s been any improvement on that front. But based on analyst estimates for their upcoming quarterly results, analysts are actually expecting that revenue for Reality Labs, which is Meta’s division that works on all these things will actually be down almost 19% compared to a year prior. So the expectations aren’t really high on that front.

Zoe Thomas: All right. That’s our reporter, Salvador Rodriguez. Thanks for joining us, Sal.

Salvador Rodriguez: Thank you for having me.

Zoe Thomas: To hear more tech stories from the Wall Street Journal, ask your Google Assistant to play WSJ Tech News Briefing podcast.

Source:

https://www.wsj.com/podcasts/google-news-update/ai-could-be-key-to-facebook-parent-metas-turnaround/b5f77b74-f347-4d6a-bfe2-2fb541ee562e?page=1