Five years ago, Facebook unveiled one of the most surprising rebrands of the decade. While the social media platform itself kept the Facebook name, the parents company, the one that also owns WhatsApp and Instagram, became Meta.
The whole thing coincided with Mark Zuckerberg declaring the metaverse “the successor to the mobile internet”, and pouring billions into the project. Zuckerberg’s vision has faced ridicule for years, but after recent developments, it’s now looking like one of the worst rebrands of all time.
Facebook’s metaverse avatars were routinely mocked (Image credit: Meta)
As reported by CBNC, the company is shutting down its VR platform Horizon Worlds in a “further pivot away from the metaverse”. In a community blog, Meta announced that the Horizon Worlds app will be taken off the Quest store at the end of March, and fully removed from VR on June 15.
Back in January, as reported by Wall Street Journal, Meta laid off 10% of its Realty Labs division, representing around 1,500 jobs. Reality Labs is the home of Meta’s AR and VR divisions.
Zuckerberg’s vision for the Metaverse has already faced ridicule from a design perspective, with rudimentary graphics and missing legs. But the laying off of over 1,000 staff from the metaverse division is a much more stark and serious sign that Zuckerberg’s 2021 vision might not be the future.
Facebook’s Meta rebrand suddenly doesn’t look so smart (Image credit: Meta)
“We said last month that we were shifting some of our investment from Metaverse toward Wearables,” a Meta spokesman told WSJ. “This is part of that effort.”
Zuckerberg once thought this was the future (Image credit: Meta/Mark Zuckerberg)
Meta might have perfectly valid reasons for shifting investment, but it can’t scrub its 2021 messaging from the history books. “The metaverse will eventually encompass work, entertainment, and everything in between,” the company announced when unveiling its rebrand. Back then, Meta announced $10B of investment into Reality Labs.
All of which is to say, with the benefit of hindsight, that rebrand isn’t looking so smart anymore. Still, to be fair, few knew in 2021 what the tech landscape would look like today – the likes of NFTs looked like they could change everything, along with the metaverse. But five years later, it looks like Facebook might have been better off changing its name to ‘AI’.
Daniel John is Design Editor at Creative Bloq. He reports on the worlds of design, branding and lifestyle tech, and has covered several industry events including Milan Design Week, OFFF Barcelona and Adobe Max in Los Angeles. He has interviewed leaders and designers at brands including Apple, Microsoft and Adobe. Daniel’s debut book of short stories and poems was published in 2018, and his comedy newsletter is a Substack Bestseller.
WASHINGTON — Meta and Google enlisted trusted children’s brands such as Sesame Street, Girl Scouts and Highlights magazine to teach kids to use technology in moderation — even as the companies designed apps that made it difficult for those same young users to unplug, public statements and internal documents show.
Backed by tens of millions of dollars from the tech giants, these organizations delivered lessons about personal responsibility to hundreds of thousands of children and parents, using colourful magazines, popular characters and catchy songs, according to public statements.
The tech giants designed apps that made it difficult for those same young users to unplug, public statements and internal documents show.Syda Productions – stock.adobe.com
The partnerships also weaken trust in decades-old institutions families have relied on for advice on raising kids, parent advocates said, at a time when the tech giants are facing down multiple lawsuits accusing them of designing addictive products that harmed youth mental health.
The first case to reach trial ended with a $6 million judgment against the two companies.
“It’s like Sesame Street teaming up with Philip Morris to teach kids how to smoke cigarettes safely,” said Rose Bronstein, whose 15-year-old son died by suicide after he was bullied online. “How is it any different?”
Meta and Google’s properties generate billions of dollars in advertising revenue from businesses marketing to minors. That economic incentive, critics say, makes it difficult for the companies to offer unbiased guidance on screen use.
“Their very business model relies on maximum time on device,” said Emily Boddy, co-lead of US Smartphone Free Childhood, a parent group that advocates against phones in schools. “Their guidance or advice can’t be neutral, and we see that it’s not.”
Corporations, ranging from soda companies to the tobacco industry, have long made donations to “trusted institutions” to improve their reputations, said Nora Kenworthy, a public health researcher at the University of Washington Bothell.
“It’s very much a reputation management strategy,” Kenworthy said.
Sponsorships extend across several brands
Reuters reviewed thousands of pages of company documents made public through lawsuits, along with company-sponsored educational videos and lessons.
The documents reveal that Meta’s strategy to partner with outside groups to promote positive messages about technology began several years ago as criticism of the apps started to proliferate.
In a 2018 draft document, internal user experience researchers deliberated how to respond to accusations that social media companies were “designing addictive products that can harm well‑being.”
Researchers proposed asking external experts to identify Facebook features that could have a negative effect on users over time.
In a 2018 draft document, internal user experience researchers deliberated how to respond to accusations that social media companies were “designing addictive products that can harm well‑being.”Davide Angelini – stock.adobe.com
Among their list of ideas, they wrote: “Form an alliance where the third party can vouch for the thoroughness and relevance of our approach for targeting the ‘addiction’ claims.” In a statement to Reuters, Meta said it did not act on that idea.
The companies did establish relationships with numerous brands. Google sponsored Sesame Street, Highlights and Girl Scouts. Meta also sponsored Girl Scouts.
Some of the materials promoted by Meta and Google do include digital safety instructions, children’s media researchers said, including reminders to set strong passwords and avoid scams.
The companies declined to say what they paid these organizations. But in a 2024 statement, Google pledged to spend at least $20 million supporting groups that promote “digital well-being,” including Highlights Magazine and Sesame Workshop.
“We prioritize the well-being of our youngest users by building industry-leading safeguards and putting families in charge of their digital experiences — any suggestion otherwise is simply wrong,” a Google spokesperson told Reuters.
Sesame Workshop said Google had no control over its digital well-being educational materials, adding in a statement that Google executives gave advice “prior to the start of content development.” Child development researchers, parents and caregivers weighed in on the materials themselves, Sesame said.
Meta said in a statement it had a limited role in designing the Girl Scout materials, but said it was proud of its work with experts in online safety. The company often works with academics to study negative use of the platform, a spokesperson said.
Highlights Magazine declined to answer specific questions about its Google partnership. Spokesperson Melanie Bay said the magazine designs products to help kids “make thoughtful choices.”
Merit badges for using tech
The Girl Scouts’ digital safety curriculum, sponsored by Meta’s Instagram, requires that girls complete age-specific lessons to earn a “digital leadership” badge.
One part of the curriculum aimed at middle-school-aged scouts instructs girls to track their screen time. Girls are then challenged to “create digital content to support a topic” they care about.
The Girl Scouts’ digital safety curriculum, sponsored by Meta’s Instagram, requires that girls complete age-specific lessons to earn a “digital leadership” badge.
Last year, Google began sponsoring its own Girl Scouts patch, called the “Be Internet Awesome Fun Patch,” tied to the company’s digital literacy curriculum. Girls learn about being kind online, using strong passwords, and keeping personal information private.
The patch, available on the Girl Scouts website, features its logo, as well as Google’s.
“It’s almost priming them to desire to get on social media once they reach the minimum age,” said Brendesha Tynes, a children’s media researcher at the University of Southern California.
Girl Scouts did not respond to multiple requests for comment.
Smartphone sleeping bags
Google also paid Highlights magazine at least $5 million. A 2024 special edition sponsored by Google includes instructions on how to make a “sleeping bag” to store devices overnight.
“Before you shut down for the night, put your device to bed,” the magazine says.
The activity makes it appear normal for Highlights readers — who range in age from six to 12 — to have smartphones at that age, seven parents who advocate for tech restrictions told Reuters after reviewing the magazine.
Google paid Highlights magazine at least $5 million.Christopher Sadowski
Google provided an extra 250,000 copies of the special Highlights edition to organizations such as Save the Children and Reading is Fundamental.
In a statement, a Google spokesperson said the company’s internet safety curriculum is “accredited and reputable,” adding that Google worked with safety organizations to design it.
One of those organizations is the Family Online Safety Institute, a non-profit that receives the majority of its revenue from tech companies, including Google. Meta is not a member.
The institute said in a statement that they reviewed the curriculum before launch.
Some consequences addressed
The lessons sponsored by Google and Meta addressed some of the apps’ effects on kids, four children’s media researchers and paediatricians told Reuters.
Meta’s sponsored Girl Scouts curriculum for middle schoolers addresses how companies take user data to sell products or “influence you online.”
A Scholastic worksheet sponsored by Google asks kids to practice what to do if they get a pop-up message that says, “You’ve won a free smartphone! Click here to get it!”
In short: Meta faces a convergence of lawsuits across the US, Australia, and the UK alleging the company knowingly profited from scam ads on Facebook and Instagram, with its own internal documents projecting that 10% of 2024 revenue, roughly $16 billion, came from fraudulent advertising. The cases span a $500 million pump-and-dump scheme, deepfake celebrity endorsements, financial professional impersonation, and cryptocurrency fraud, while leaked internal assessments show Meta calculated that scam revenue would exceed the cost of any regulatory settlement.
Meta is facing a convergence of lawsuits, class actions, and regulatory investigations over scam advertisements on Facebook and Instagram that, according to the company’s own internal projections, generated roughly $16 billion in revenue in 2024, approximately 10% of Meta’s total advertising income. The legal actions span the United States, Australia, and the United Kingdom, and collectively allege that Meta knowingly profited from fraudulent ads including AI-generated deepfake celebrity endorsements, pump-and-dump stock schemes, fake investment platforms, and unauthorised impersonation of financial professionals, while maintaining ad moderation systems that were structurally inadequate to prevent the fraud and, in some cases, deliberately weakened to protect revenue.
The most financially significant case, filed in February in the US District Court for the Northern District of California, alleges that Meta facilitated a pump-and-dump scheme involving Jayud Global Logistics, a Chinese stock listed on Nasdaq. According to the complaint, scammers acquired 50 million shares at discounted prices in December 2024, then used targeted Facebook and Instagram ads to drive the share price to nearly $8 before dumping their positions in April 2025. Consumer losses exceeded $500 million. A California federal judge dismissed the class action on 25 March, ruling that the plaintiffs had not sufficiently alleged Meta “co-created” the ads, though the dismissal appeared to be without prejudice.
The pattern across jurisdictions
A separate class action, filed by Scott+Scott on behalf of financial professionals John Suddeth and Sara Perkins, alleges that Meta allowed scammers to use their names, images, voices, and professional personas in paid advertisements, causing client diversion, reputational harm, and regulatory inquiries. A bipartisan coalition of US state attorneys general had warned Meta in June 2025 that impersonation ads and fraudulent WhatsApp investment groups were being used for widespread fraud. According to the complaint, materially identical impersonation ads continued running after the warning.
In December, the US Virgin Islands attorney general sued Meta in Superior Court, alleging the company “knowingly profited” from scam ads and “charged fraudsters extra for the right to advertise scams” rather than removing them. The Virgin Islands suit joined actions by 42 other state attorneys general who have taken Meta to court, primarily over child safety but with increasing overlap with advertising fraud. In April, New York attorney general Letitia James issued an investor alert specifically about investment scams on Meta platforms.
In Australia, the Competition and Consumer Commission has been pursuing Meta in federal court since March 2022 over cryptocurrency scam ads that used the likenesses of businessman Dick Smith, television presenter David Koch, and former New South Wales premier Mike Baird. A single victim cited in the complaint lost more than A$650,000. Meta failed to get the case dismissed in 2023. In the United Kingdom, the Financial Conduct Authority found 1,052 illegal financial advertisements on Meta platforms in a single week in November 2025. A leading UK bank found that 80% of its fraud cases originated on Meta’s platforms, with Facebook Marketplace accounting for 60% of purchase fraud, Instagram responsible for 67% of investment fraud, and WhatsApp impersonation scams up 300% year on year. Meta’s platforms account for 61% of all authorised push payment scams in the UK, according to UK Finance, with criminals stealing £485.2 million.
The $16 billion question
The scale of the problem is defined by Meta’s own internal documents. A Reuters investigation published in November 2025 revealed that Meta projected 10% of its 2024 global revenue, roughly $16 billion, derived from scam and fraud-related advertising. The company served an estimated 15 billion “higher risk” scam ads per day. Nineteen percent of Meta’s ad revenue from China, approximately $3 billion, was linked to scams. Internal documents showed that when enforcement staff proposed shutting down fraudulent accounts, Meta sought assurance that “growth teams would not object given the revenue impact.” A subsequent Reuters report in January found that Meta had developed an internal “playbook” to neutralise regulators and manipulated its ad library to make scam ads harder to find.
Rob Leathern, Meta’s former senior director of product management who led business integrity operations, said of the findings: “The levels that you’re talking about are not defensible.” Meta described the Reuters projections as “a rough and overly-inclusive estimate” and said the documents presented “a selective view that distorts Meta’s approach to fraud and scams.”
The economics are straightforward. Implementing universal advertiser verification would cost Meta approximately $2 billion and reduce revenue by up to 4.8%. Internal assessments reportedly noted that “revenue from risky ads would almost certainly exceed the cost of any regulatory settlement,” a calculation that treats fines as a cost of doing business rather than a deterrent.
The deepfake dimension
AI-generated deepfakes have become central to the scam ad ecosystem. Deepfake fraud attempts have surged by 3,000% as generative AI tools have become cheaper and more accessible, enabling scammers to create convincing fake video endorsements at scale. Martin Lewis, the UK’s most prominent personal finance campaigner, was targeted with a deepfake video promoting a “Quantum AI” investment scheme. Deepfakes of Donald Trump, Elon Musk, Alexandria Ocasio-Cortez, and Bernie Sanders were used to promote fake government benefit schemes. In Brazil, AI-altered images and voices of prominent physicians promoted fraudulent healthcare products.
The Tech Transparency Project identified 63 scam advertisers responsible for more than 150,600 political ads and $49 million in lifetime spending on Meta. During a 90-day period in mid-2025, at least 45 scam advertisers spent over $18 million. Meta says it protects images of 500,000 celebrities and public figures through automated detection and is testing facial recognition technology to compare faces in suspected scam ads against public figures’ profile pictures. EU lawmakers have agreed to ban AI-generated non-consensual deepfakes through amendments to the AI Act, signalling increasing regulatory appetite to legislate against synthetic media that platforms have failed to police.
What Meta says it is doing
Meta recently rolled out new scam detection tools across Facebook, Instagram, WhatsApp, and Messenger. The company says it removed 159 million scam ads and took down 10.9 million accounts linked to scam operations in 2025, with 92% of scam ads caught proactively before any user report. It disabled 150,000 accounts associated with Southeast Asian scam centre networks and partnered with the Royal Thai Police in disruption operations that led to 21 arrests. Meta is targeting 90% of ad revenue from verified advertisers by the end of 2026, up from 70%. In February, it filed its own lawsuits against scam advertisers in Brazil, China, and Vietnam, and sent cease-and-desist letters to eight former Meta Business Partners offering “un-ban” services to fraudulent advertisers.
The gap between Meta’s enforcement claims and the data in its own internal documents is the through line connecting every lawsuit. The company says it catches 92% of scam ads proactively. Its own projections estimated $16 billion in scam-related revenue in a single year. It removed 159 million scam ads. It served 15 billion higher-risk ads per day. It is investing in facial recognition to detect deepfakes. Its internal assessments concluded that scam revenue would exceed the cost of any regulatory settlement. The numbers do not cohere into a story of a company that failed to notice the problem. They describe a company that noticed the problem, measured it, calculated the cost of fixing it against the cost of not fixing it, and chose the option that preserved revenue. The lawsuits are, in that sense, not about whether Meta knew. They are about what it did with what it knew.
When news broke Tuesday morning that Meta bought Moltbook, the social network for AI agents, it may have left some people scratching their heads. What on earth would Meta — an ad-supported company — want with a social network where the users are bots? Bots, after all, are not the target audience of brand marketers and advertisers.
Meta isn’t saying much. Its only official comment was a brief statement that the Moltbook team was joining Meta Superintelligence Labs, which would open up “new ways for AI agents to work with people and businesses.”
Reading between the lines, this was an acqui-hire. A network built for bots isn’t exactly a natural home for brand advertising — even if Moltbook was never entirely non-human. What Meta really wanted was the talent behind it — people who are having fun brainstorming and experimenting with AI agent ecosystems. And that, counterintuitively, could be a boon for its advertising business.
As Meta CEO Mark Zuckerberg said last year, he believes in a future where “every business will soon have a business AI, just like they have an email address, social media account, and website.” On an agentic web, one where AI systems act independently on users’ behalf, AI agents could interact with each other, doing things like buying ads, making bookings, and responding to customers.
AI is also being used to generate ad creative and tailor its output based on who’s viewing it. AI systems could also manage product pricing or generate personalized offers.
On the consumer side, agents could be used to find the best prices and deals, manage bookings, and shop for products. In somelimitedcases, agents can already check out and pay on consumers’ behalf. (Agentic commerce is still in its early days, and these systems don’t always work as well as advertised. But the market has been moving fast, and improvements seem likely soon enough.)
As Facebook once built the “friend graph” — a network defined by social connections between people, where every individual is a node — an agentic web could benefit from an “agent graph,” a system that maps out how various agents are connected and what actions they can take on each other’s behalf.
For an agentic web where businesses’ agents and consumers’ agents can work together, though, the agents first need to be able to find each other, connect, and coordinate their activities. As Facebook once built the “friend graph” — a network defined by social connections between people, where every individual is a node — an agentic web could benefit from an “agent graph,” a system that maps out how various agents are connected and what actions they can take on each other’s behalf. This could span areas like travel, online shopping, media and research, productivity tools, and more.
This, too, could be where advertising slots in. Today, humans view and click on ads when they see something of interest, but on an agentic web where agents are shopping on users’ behalf, ads might look quite different. Instead of influencing a human to buy a product, a business’s agent may need to negotiate directly with a consumer’s agent to make the sale.
Maybe the consumer wants to buy that shirt or that lipstick, but only in a certain colour and at a certain price. Maybe the systems become so complex that these considerations go beyond product and price — perhaps the consumer prefers to support small businesses, or shops only with eco-friendly companies. Maybe the consumer only buys items when they’re on sale or purchases generic versions if the ingredients are the same. And so on.
In that case, it’s not just a matter of connecting the AI agents but also ranking products by whichever one best fits that individual customer’s needs. If Meta could capitalize on that market — AI at the orchestration layer, meaning the system decides which agents talk to each other and in what order — it could potentially expand its ads business into entirely new territory.
This all depends on whether consumers actually embrace the agentic web, or ever trust AI enough to let it act on their behalf. But the very existence of OpenClaw, the personal AI assistant that populated Moltbook with content, suggests that at least some people are already leaning into autonomous AI agents.
Of course, there’s another possible reason Meta bought Moltbook. The company lost the acqui-hire of OpenClaw’s creator, Peter Steinberger, to rival OpenAI, so it went after Moltbook, the platform Steinberger’s tool helped build, instead. Petty? Maybe. But it kept Meta’s Superintelligence Labs in the news.
Meta has announced a heap of new ad updates, primarily focused on retailers and those using its automated Advantage+ campaigns.
And there are a lot of niche use cases within these new updates, which could apply to your business.
The first update is “Advantage+ creative optimizations”, which will automatically optimize your video ads for viewing on Reels, or the mobile Facebook and Instagram apps with 9:16 ratio.
That will help more brands tap into the popularity of the various Meta video formats, with Reels being the key focus.
“Reels and video on our apps continues to grow as daily watch times across all video types grew over 25% year-over-year in Q4. In fact people now reshare Reels 3.5 billion times every day.”
The new process will also enable advertisers to dynamically create multiple variations of an ad, so the system then has more options to display to users, depending on what they respond best to.
Meta’s also updating its Advantage+ catalogue ads, with the added capacity to import and use branded videos or customer demonstration videos, instead of just static images.
Advantage catalogue ads, which Meta first launched in beta testing last year, provide personalized recommendations to users, based on what Meta’s system detects that each will be most interested in, and this new process will provide more capacity to showcase relevant products within the display.
Meta’s will also now enable brands to upload a “hero” image in the centre of their catalogue ads, which Meta’s AI will then use to show people the best products from their catalogue to drive performance.
Meta’s also adding more eCommerce ad options, with users of Magento and Salesforce Commerce Cloud now able to create Shops ads within their management systems. Meta’s also integrating its Shops ads and branded content ads (now called “Partnership ads”), which will enable direct purchasing from collaborative campaigns.
And there’s also new elements in Reminder Ads on Instagram:
“Now, advertisers can include external links to a new product or sale in their Reminder ads to help turn a person’s interest into a purchase. This summer, we’ll also give advertisers ways to notify people when an event starts and before it ends.”
Meta’s also looking to expand Reminder ads to Reels in the coming months.
There’s also new Promo Codes promotions on Facebook and Instagram, as well as ads with product tags:
“In March, we’ll bring ads with product tags to Facebook Feed (currently Instagram only), and in April, we’ll launch the global availability of ads with product tags to all businesses, whether or not they maintain a Shop.”
Meta’s also updating its Collaborative ads offering, to provide more analytics on performance, while it’s also testing the ability for advertisers to use Collaborative ads with Advantage+ shopping campaigns.
Finally, Meta’s also working on Advantage+ Catalogue ads with omnichannel brand and product level reporting as a new managed service solution “to help RMNs prove that ads on Meta platforms drove sales online and in-store”.
So yeah, a heap of updates, all with varying levels of applicability and use. And while there may not be some huge, headline change that will get the most attention, there’s a lot of value for specific brands within these changes.
Between a behemoth copycat and a looming ban, TikTok is being attacked on all fronts.
When Instagram releases a new feature that is a direct copy of another app, its users fear the worst.
In a 2022 essay for Digital Trends, writer Cristina Alexander lamented the “TikTok-ification of Instagram” because it “takes away the type of content people love most about the platform: photos from friends and family, as well as content based on their interests.”
“And it’s something I’m just about fed up with,” Alexander wrote.
But the doomsaying rarely lasts forever. Alexander joins the ranks of Kylie Jenner, Kim Kardashian, and a whole host of regular users — including myself — who have fallen into the cycle of hating it when Instagram makes a copycat change and then, after a few months, come around to it.
Like it or hate it, Instagram’s copycat strategy works — and its dedication to stealing features from other apps is helping to fuel its ability to overtake TikTok.
Think of Instagram like Kirby in Super Smash Bros. He’s a formidable foe on his own, but it’s using his Copy Ability by swallowing his enemies and using their own powers against them that makes him so powerful. Instead of finding and using power ups or prioritizing his abilities, Kirby uses his enemies as his own, personal power ups. Instagram — and other Meta-owned apps — swallow their enemies, take on their features, and use them to win. Instagram used this strategy to remove Snapchat from its list of significant competitors, and TikTok is next.
For the first time since 2020, Instagram overtook TikTok in new app downloads in 2023, according to data from market intelligence firm Sensor Tower reported by the Financial Times, making it the most downloaded app in the world. In 2023, Instagram downloads grew 20 percent in comparison to TikTok’s 4 percent.
This comes after Instagram launched Reels, a TikTok-esque feature that was originally panned by its user base but has now become a mainstay on the app. And it might be the inclusion of Reels that has helped launch the platform back to the top.
“Instagram has outperformed TikTok in adoption over the past few years, driven by the popularity of its Reels feature along with legacy social media features and functions,” Abraham Yousef, a senior insights manager at Sensor Tower, told the Financial Times.
Instagram’s successful copycat strategy might be the reason it is succeeding, but TikTok is facing a battle at all fronts.
President Joe Biden said that if Congress passes the “Protecting Americans from Foreign Adversary Controlled Applications Act” — which would ban TikTok and all other apps based in China, North Korea, Russia, and Iran from U.S. app stores — he’ll sign it into law. Lawmakers argue that TikTok user data for U.S. citizens could be accessed by the Chinese state, but TikTok has consistently denied that claim.
All the while, TikTok is becoming increasingly less fun and more focused on ecommerce. With the emergence of TikTok shop, it feels like every other video on the For You Page is a promoted or sponsored post. The TikTok experience is changing, and it might not be for the better.
Just because fewer people are downloading the app, and many more are complaining about their experience on it, doesn’t mean TikTok is fully failing, though. The app has higher engagement than its rivals, with users spending an average of 95 minutes on TikTok in comparison to 62 minutes on Instagram, 30 minutes on X, and 19 minutes on Snapchat, according to the Financial Times report.
We’ll have to wait and see what a TikTok ban will look like, but one thing is certain — even if the app isn’t banned in the U.S., the fight for users won’t be over.
Feature Image Credit: Photo by Matt Cardy/Getty Images
Sheryl Sandberg, who helped to transform Facebook from a tech startup into a digital advertising empire, will step down from the board of Meta, Facebook’s parent company.
“With a heart filled with gratitude and a mind filled with memories, I let the Meta board know that I will not stand for re-election this May,” Sandberg wrote in a Facebook post.
Sandberg left Google to join Facebook in 2008, four years before the company went public. As the No. 2 executive at Meta under CEO Mark Zuckerberg, Sandberg also took a lot of heat for some of its biggest missteps.
She stepped down as chief operating officer of Meta in 2022 but remained a member of the company board. She had served as COO of Facebook, and then Meta, for 14 and a half years and as a board member for 12 years.
“Under Mark’s leadership, Javi Olivan, Justin Osofsky, Nicola Mendelsohn, and their teams have proven beyond a doubt that the Meta business is strong and well-positioned for the future, so this feels like the right time to step away,” Sandberg wrote.
Sandberg said she will continue to serve as an advisor to the company.
Last year Sandberg announced that she was launching a girls leadership program through her foundation to respond to what she calls stubborn gender inequities. The girls leadership program includes a middle-school curriculum as well as resources for adults.
Lean In is a project of the Sandberg Goldberg Bernthal Family Foundation, the private foundation Sandberg started with her late husband, Dave Goldberg.
Meta announced it’s rolling out its first generative AI features for advertisers, allowing them to use AI to create backgrounds, expand images and generate multiple versions of ad text based on their original copy. The launch of the new tools follows the company’s Meta Connect event last week where the social media giant debuted its Quest 3 mixed-reality headset and a host of other generative AI products, including stickers and editing tools, as well as AI-powered smart glasses.
In the case of AI tools for the ad industry, the new products may not be as wild as the celebrity AIs that let you chat with virtual versions of people like MrBeast or Paris Hilton, but they showcase how Meta believes generative AI can assist the brands and businesses that are responsible for delivering the majority of Meta’s revenue.
The first among the trio of new features allows an advertiser to customize their creative assets by generating multiple different backgrounds to change the look of their product images. This is similar to the technology that Meta used to create the consumer-facing tool Backdrop, which allows users to change the scene or the background of their image by using prompts. However, in the ad toolkit, the backgrounds are generated for the advertiser based on their original product images and will tend to be “simple backgrounds with colours and patterns,” Meta explains. The feature is available to those advertisers using the company’s Advantage+ catalogue to create their sales ads.
Another feature, image expansion, allows advertisers to adjust their assets to fit different aspect ratios required across various products, like Feed or Reels, for example. Also available to Advantage+ creative in Meta’s Mads Manager, the AI feature would allow advertisers to spend less time repurposing their creative assets, including images and video, for different surfaces, Meta claims.
With the text variations feature in Meta Ads Manager, the AI can generate up to six different variations of text based on the advertiser’s original copy. These variations can highlight specific keywords and input phrases the advertiser wants to emphasize, and advertisers can edit the generated output or simply choose the best one or ones that fit their goals. During the campaign, Meta can also display different combinations of text to different people to see which ones drive better responses. However, Meta won’t showcase the performance details for each specific text variation, it says, as the reporting is currently based on a single ad. However, the more options the advertiser selects to run, the more opportunities they’ll have to improve their ad performance, Meta informs them.
Meta says it’s already tested these AI features with a small but diverse set of advertisers earlier this year, and their early results indicate that generative AI will save them five or more hours per week, or a total of one month per year. However, the company admits that there’s still work ahead to better customize the generative AI output to match each advertiser’s style.
In addition, Meta says there are more AI features to come, noting it’s working on new ways to generate ad copy to highlight selling points or generative backgrounds with tailored themes. Plus, as it announced at Meta Connect, businesses will be able to use AI for messaging on WhatsApp and Messenger to chat with customers for e-commerce, engagement and support.
Meta is reducing its engagement with news publishers, focusing less on current affairs on its platforms.
The company launched a text-based app, Threads, that prioritizes non-news content.
Meta is in conflict with Canada’s government over legislation requiring platforms to pay for publishers’ content.
Tensions are escalating as Meta, the parent company of Facebook, Whatsapp, Threads, and Instagram grows increasingly distant from news publishers, sparking widespread concern.
This move comes amidst a shift in strategy where the technology titan has been giving less attention to politics and current affairs on its platforms, while simultaneously shrugging off governmental calls for increased payments to media outlets.
Meta’s growing reluctance: A strategic move
In a pivotal turn, Meta has distanced itself from the traditional news sector, despite years of appeasing key publishers through the funding of non-profit journalism initiatives and forging partnerships with entities like Rupert Murdoch’s News Corp.
This alteration in stance manifests in Meta’s latest product release – Threads, a text-based application designed to rival Twitter. In less than a week, Threads managed to draw in an astounding 100 million users, thanks to its integration with the globally popular Instagram platform.
Much like Instagram, Threads prioritizes content from creators and friends over hard news or political stories. Adam Mosseri, the head of Instagram, has firmly declared the platform’s intent to avoid promoting news content.
In a controversial move, Meta has decided to exclude news from its feed in Canada, as new legislation demanding platforms pay for content from publishers and broadcasters comes into effect.
This law was formulated to uplift smaller news organizations with limited bargaining power. However, the law has been met with resistance not just from Meta, but also Google, which threatens to impose a news blackout in Canada.
The corporate tussle has elicited backlash from a host of advertisers in Canada, some of which are threatening to withdraw their advertisements. The implications for Meta are significant, given that Canada contributed approximately $3 billion to the company’s $117 billion annual revenues in 2022.
A history of friction
Historically, Meta has made attempts to ally with publishers through various initiatives, such as deals for content to be featured on Facebook’s News Tab product.
However, the senior leadership at Meta has concluded that the company’s interests conflict with those of the news industry. This stems from the notion that the growth of the company’s digital advertising business is perceived as a contributing factor to the global revenue decline experienced by newspaper groups.
Additionally, Meta’s internal research has revealed that users gravitate more towards short-form videos and content from influencers rather than news and political content. Consequently, the tech giant has reduced the presence of political content in users’ feeds since 2021.
Despite its ongoing withdrawal from the news industry, the ramifications of Meta’s actions are far-reaching.
With allegations that the company’s inadequate moderation of its applications fuelled discord surrounding the election of former US president Donald Trump, as well as the 2021 Capitol building riots, the technology behemoth is treading on thin ice.
As a result, industry insiders argue that Meta will eventually suffer from the escalating rift with news publishers. The absence of reliable news sharing could potentially isolate the firm from real-world happenings, leading to the question of whether its strategy will prove sustainable in the long term.
As the tension unfolds between Meta and news publishers, the future of news content on social media platforms remains uncertain.
However, one thing is clear: the technology titan’s standoff with news organizations and governments alike is set to redefine the relationship between social media and the world of journalism.
Jai Hamid is an enthusiastic writer whose current area of interest is the blockchain sector. Whenever she is not reading or writing, you can find her tending to her plants in the garden. She strongly believes that crypto is going to transform the world for the better.
The company hopes that making LLaMA 2 open source might give it the edge over rivals like OpenAI.
Meta is going all in on open-source AI. The company is today unveiling LLaMA 2, its first large language model that’s available for anyone to use—for free.
Since OpenAI released its hugely popular AI chatbot ChatGPT last November, tech companies have been racing to release models in hopes of overthrowing its supremacy. Meta has been in the slow lane. In February when competitors Microsoft and Google announced their AI chatbots, Meta rolled out the first, smaller version of LLaMA, restricted to researchers. But it hopes that releasing LLaMA 2, and making it free for anyone to build commercial products on top of, will help it catch up.
The company is actually releasing a suite of AI models, which include versions of LLaMA 2 in different sizes, as well as a version of the AI model that people can build into a chatbot, similar to ChatGPT. Unlike ChatGPT, which people can access through OpenAI’s website, the model must be downloaded from Meta’s launch partners Microsoft Azure, Amazon Web Services, and Hugging Face.
“This benefits the entire AI community and gives people options to go with closed-source approaches or open-source approaches for whatever suits their particular application,” says Ahmad Al-Dahle, a vice president at Meta who is leading the company’s generative AI work. “This is a really, really big moment for us.”
But many caveats still remain. Meta is not releasing information about the data set that it used to train LLaMA 2 and cannot guarantee that it didn’t include copyrighted works or personal data, according to a company research paper shared exclusively with MIT Technology Review. LLaMA 2 also has the same problems that plague all large language models: a propensity to produce falsehoods and offensive language.
The idea, Al-Dahle says, is that by releasing the model into the wild and letting developers and companies tinker with it, Meta will learn important lessons about how to make its models safer, less biased, and more efficient.
A powerful open-source model like LLaMA 2 poses a considerable threat to OpenAI, says Percy Liang, director of Stanford’s Center for Research on Foundation Models. Liang was part of the team of researchers who developed Alpaca, an open-source competitor to GPT-3, an earlier version of OpenAI’s language model.
“LLaMA 2 isn’t GPT-4,” says Liang. And in its research paper, Meta admits there is still a large gap in performance between LLaMA 2 and GPT-4, which is now OpenAI’s state-of-the-art AI language model. “But for many use cases, you don’t need GPT-4,” he adds.
A more customizable and transparent model, such as LLaMA 2, might help companies create products and services faster than a big, sophisticated proprietary model, he says.
“To have LLaMA 2 become the leading open-source alternative to OpenAI would be a huge win for Meta,” says Steve Weber, a professor at the University of California, Berkeley.
Under the hood
Getting LLaMA 2 ready to launch required a lot of tweaking to make the model safer and less likely to spew toxic falsehoods than its predecessor, Al-Dahle says.
Meta has plenty of past gaffes to learn from. Its language model for science, Galactica, was taken offline after only three days, and its previous LlaMA model, which was meant only for research purposes, was leaked online, sparking criticism from politicians who questioned whether Meta was taking proper account of the risks associated with AI language models, such as disinformation and harassment.
To mitigate the risk of repeating these mistakes, Meta applied a mix of different machine learning techniques aimed at improving helpfulness and safety.
Meta’s approach to training LLaMA 2 had more steps than usual for generative AI models, says Sasha Luccioni, a researcher at AI startup Hugging Face.
The model was trained on 40% more data than its predecessor. Al-Dahle says there were two sources of training data: data that was scraped online, and a data set fine-tuned and tweaked according to feedback from human annotators to behave in a more desirable way. The company says it did not use Meta user data in LLaMA 2, and excluded data from sites it knew had lots of personal information.
Despite that, LLaMA 2 still spews offensive, harmful, and otherwise problematic language, just like rival models. Meta says it did not remove toxic data from the data set, because leaving it in might help LLaMA 2 detect hate speech better, and removing it could risk accidentally filtering out some demographic groups.
Nevertheless, Meta’s commitment to openness is exciting, says Luccioni, because it allows researchers like herself to study AI models’ biases, ethics, and efficiency properly.
The fact that LLaMA 2 is an open-source model will also allow external researchers and developers to probe it for security flaws, which will make it safer than proprietary models, Al-Dahle says.
Liang agrees. “I’m very excited to try things out and I think it will be beneficial for the community,” he says.
