The Washington PostDemocracy Dies in Darkness

Facebook pivoted to the metaverse. Now it wants to show off its AI.

Meta released new generative AI advertising tools, as its competitors push forward with bolder products

An avatar of Mark Zuckerberg, chief executive officer of Meta Platforms, speaks during the virtual Meta Connect event in New York in Oct. 2022. (Michael Nagle/Bloomberg)
6 min

Last month, internal chatter about Meta’s investments in artificial intelligence reached such a fever pitch that some employees began to wonder if the company was making a pivot away from virtual reality.

During a companywide meeting in April with top executives, employees up-voted a question on the subject: Do we still care about the metaverse? Workers were told the answer was yes.

That Meta workers even had to ask was notable. It was less than two years since Mark Zuckerberg had renamed the company he founded as Facebook to Meta, staking its future on the immersive digital realms called “the metaverse.”

Now, Meta’s chief executive officer is trying to project the social media giant as a serious player in artificial intelligence, as an arms race to command the burgeoning technology is captivating Silicon Valley. While Microsoft, Google and a string of start-ups have pushed out chatbots and other splashy innovations using generative artificial intelligence, a form of the technology that can produce humanlike original content, Meta appears to be playing catch-up.

Two of its generative AI projects including Galatica, a large-language model for synthesizing scientific research, and its BlenderBot 3 all-purpose chatbot, faced negative publicity when the models spewed inaccurate and hateful rhetoric. The company quickly paused the public demo of Galatica while BlenderBot failed to gain much traction. Since then, the company has touted more mundane innovations, including advertising tools.

Zuckerberg endorsed artificial intelligence as a driving force behind Meta’s revenue growth last quarter, after nearly a year of dismal financial performance.

“Our AI work is driving good results across our apps and business,” Zuckerberg said in a statement last month at the time of the earnings release.

Meta starts cutting 10,000 workers in second round of layoffs

The Meta CEO has forcefully denied its newfound attention on AI signals the company is less committed to its larger metaverse bet. In fact, Zuckerberg has argued AI is foundational to the metaverse, and will be a critical tool to create more dynamic and accessible virtual reality-powered experiences. For instance, less tech-savvy users will be able to use generative AI to create their own new worlds in Meta’s virtual reality-powered apps.

But both Wall Street and workers will be watching how Zuckerberg prioritizes investing in both technologies as he seeks to shore up the company’s financial performance and employee morale.

“It’s crucial that Meta outlines the strategy, how they’re going to monetize it and flex the muscles to show that they also are a key AI player,” said Wedbush Securities analyst Dan Ives. They should demonstrate “they’re not sitting at the kids’ table with Microsoft, Google and Apple at the adults table.”

The rhetoric about AI has confused some employees and investors about the company’s core focus, as a string of news articles and analysts mulling whether the metaverse is dead and Zuckerberg is walking away from his multibillion-dollar investment.

Some workers expressed frustration that the company’s direction is convoluted, as it endorses AI while cutting back on jobs and other projects, according to three people familiar with the matter who spoke on the condition of anonymity to speak about internal matters. Other workers rushed to receive an internal alert for AI-related jobs, a way to safely position themselves for the company’s next big bet, one of the people said.

How Mark Zuckerberg broke Meta’s workforce

On Thursday, the company announced it was launching AI Sandbox, which enables marketers to use AI to create more text and visual options for ads. Meta executives said the product will help marketers more effectively target their ads to different types of consumers, by experimenting with extremely subtle changes to copy, pictures and backgrounds.

For example, an ad might get more traction with women under 35 if the company’s logo is depicted on top of a cityscape with blurry lights rather than a snowy mountain. Meta also announced Meta Lattice, a new artificial-intelligence-driven model to improve the performance of its ads on its networks.

But these products are far more modest than those championed by its competitors in recent months. On Wednesday, Google announced it would start answering some search queries directly by generating its own results, drawing from sources around the web, rather than linking and summarizing from other websites as it has done for two decades.

That strategic shift could transform users’ experience of the internet while upending publishers and content creators who depend on Google for traffic. Google’s announcement followed pressure from investors who urged the company to catch-up with Microsoft, which had already incorporated ChatGPT into its own search engine, Bing.

Google is changing the way we search with AI. It could upend the web.

Meta appears to be taking a more reserved path.

In February, the company announced it was forming a new product group to “turbocharge” its use of generative AI. The group, led by former Apple executive Ahmed Al-Dahle, aims to bring together key teams from research and consumer-focused groups to create new products, according to the company. Zuckerberg has said he expects to build generative AI-powered chat experiences in WhatsApp and Messenger as well as innovations in business messaging and customer support.

On the research front, Meta has long been a player in artificial intelligence. The company’s laboratory stands out in the industry for publicly releasing a lot of its research. Research projects include technology using AI to animate children’s drawings, decode speech from imagery of brain activity, and enable real-time translation of most languages.

Some think Meta’s new focus on AI could help overcome threats to its ad-based business model. Increasing competition in the social media market from upstarts such as TikTok and new privacy rules from Apple have hurt the digital advertising market.

Meta begins fresh round of job cuts among highly skilled staff

Zuckerberg credited AI last month for driving people to spend more time on Instagram, as the company promotes social short-form videos called Reels. And he reminded investors during an earnings call that artificial intelligence is powering how the company recommends content to users, delivers ads to consumers and roots out rule-breaking and offensive content — systems the company has “been working on for many, many years.”

But Meta executives in recent weeks have sought to push back on claims that their investment strategy has dramatically shifted.

“A narrative has developed that we’re somehow moving away from focusing on the metaverse vision, so I just want to say upfront that that’s not accurate,” Zuckerberg said last month. “We’ve been focusing on both AI and the metaverse for years now, and we will continue to focus on both.”