Zuckerberg was just three minutes into his keynote speech at the company's annual Meta Connect product event last month when his tone turned almost wistful.

"Building is an act of love," he said, glancing down at the ground with his hands in his pockets. "It's how we impart what we believe, and we pour our hearts and our souls into what we make." Speaking to hundreds of analysts and developers, Zuckerberg proceeded to share news about Meta's latest AI products: its new agentic chatbot Muse, a slew of smart glasses, as well as a Tamagotchi-like AI gadget users can attach to their wrists.

It was a conference focused on new technologies - but still, it's perhaps notable that during a presentation lasting nearly an hour, Zuckerberg did not mention two of his most popular products, Instagram and Facebook, the social media platforms that have helped Meta become a $2tn company.

For much of the past year, Meta has been embroiled in a series of legal actions, accused of deliberately designing products that are addictive to young users. That has led to the release of internal emails, corporate documents and whistleblower testimony that personal injury lawyers and prosecutors across the country have wielded to make their case.

And Zuckerberg is unpopular: last year, a Pew Research Center poll found that two-thirds of Americans had an unfavourable opinion of him.

His reputation is unlikely to be helped by The Social Reckoning, a film by Aaron Sorkin released this week. It's about whistleblower claims that Meta executives were aware of their products harming young people. Succession actor Jeremy Strong plays Zuckerberg as a "full-on villain", according to a review in the Hollywood Reporter, "stoically arguing that even the slightest gesture of responsibility is anathema where money is concerned".

Some speculate that this is Meta's "Big Tobacco" moment, with a consensus forming against the company, and the social media industry as a whole.

But the truth seems to be more complicated. Despite months of bad press, in the real world Meta's products are doing better than ever. Use of its apps continues to rise, and users are already flocking to its new AI assistant.

It raises the question: can anything really clip Meta's wings? Or is it now something of a 'Teflon' company, with scandals simply bouncing off it? It's certainly been a year of bad headlines for Meta.

Last summer, Reuters revealed that Meta allowed its chatbots to "engage a child in conversations that are romantic or sensual", and provide false medical information to users. Meta has revised those policies and said that such responses should never have been allowed.

And this year, bills have been mounting. In March, a jury gave $6m (£4.5m) to a 20-year-old California woman who claimed she suffered mental health harms related to her use of Meta's Instagram and Google's YouTube.

Around the same time, New Mexico became the first state to succeed in a lawsuit against Meta over child safety issues. A jury found the company had failed to warn the public about the dangers its platforms posed to kids. Meta was fined a total of $942m. The judge called Meta a "public nuisance" akin to air pollution. Meta says it disagrees with the verdicts and is appealing both cases.

In May, Meta was among the social companies that avoided a trial by settling a case brought by a Kentucky school district that claimed their product design had led to a youth mental health crisis. Meta reportedly agreed to pay $9m, the most of any defendant.

And this summer, Meta struck a high-profile $18bn settlement with 48 US states, plus the District of Columbia and three US territories. Meta denied wrongdoing as part of the settlement, and during the five-day trial it maintained that it had worked extensively over the years to make its platforms safer for young users, pouring resources into testing and research. As part of the deal, the firm promised to enact two-hour daily time limits for young users, along with night-time blocks on use, muted notifications during school hours, and other changes.

The company faces other lawsuits too, including one starting later this month in Los Angeles, also related to alleged social media addiction among young users.

Meanwhile, several countries have moved to ban social media for children entirely, starting with Australia last December. Social media companies are broadly disliked, even while their products remain wildly popular. A Reuters/Ipsos survey suggests that 85% of Americans think social media can be addictive for children, and 61% support greater government oversight.

To some, the trust issue feels existential "Once the public has stopped trusting you, it's kind of a downward spiral," says associate professor Alison Taylor of New York University's Stern School of Business. "It feels like… the trust deficit is so big, I just don't know that they're going to be able to recover." But despite those high-profile cases, the company reported that the number of people using some of its apps (like Instagram and WhatsApp) has risen 3% year on year and this year's second quarter revenue was up 28% compared with the same period in 2025.

While it has agreed to temper its pursuit of the teen market (which many brands eagerly cultivate to build their future user base), the company has said that teens accounted for less than 1% of its revenue anyway. Competitors TikTok and Google's YouTube are more popular with young users.

And the company's settlement with the 48 states was widely considered a massive win for Meta, because it prevented some Meta bosses (including Zuckerberg) from having to give testimony, and stopped the release of further internal documents, while allowing the company to avoid admitting wrongdoing. The company's stock price surged 4% immediately following the announcement, and is up more than 25% over the last month.

One view popular among critics is that the company has tested consumer tolerance for bad behaviour and until now, its business has mostly managed to withstand the blowback. In other words, people will continue to use the products even if they don't like the company.

"I think they're in a remarkably good position that I didn't expect them to be in," says Patrick Moorhead, founder and CEO of Moor Insights & Strategy.

However, Meta's move into AI depends partly on rebuilding public trust.

Last month, the company launched its new AI personal assistant app, Muse. Crucially, Muse is an AI agent, meaning it can carry out tasks in multiple stages and work on something over several hours (in contrast to chatbots like ChatGPT, which tend to do one task at a time).

It's the first AI agent from a major tech firm to become publicly available. A New York Times tech reporter who used the app for two weeks says he was "blown away"; the app called his dental insurer on his behalf, ordered him groceries and tracked his credit card spending in a spreadsheet.

But Muse works best when users hand over some of their most personal information to the company, including credit card information, purchase histories and access to email inboxes.

It's a big ask for a company whose reputation on privacy has been sullied, starting a decade ago with the Cambridge Analytica scandal, which followed the revelation that the data of millions of Facebook users was breached and used by a political consulting firm.

Just two days after Zuckerberg's keynote at Meta Connect, another New Mexico jury found Meta lied about how Facebook used personal information. (The company has said it disagrees with the verdict and will defend itself against what it says are efforts to distort its record.) So are customers willing to hand over that data for the convenience of using Muse?

Some users say yes. Since its launch one month ago, Muse already has more than five million downloads and more than three million weekly active users - outpacing the adoption of ChatGPT when that was first introduced, in the North American market.

Zuckerberg has laid out a vision in which Muse couples with an ever-expanding line-up of Meta AI smart glasses, enabling users to summon hands-free what he calls "personal superintelligence".

Meta clearly has big plans for its smart glasses. "I think they're making an investment to position themselves as a category leader," says Kate Winick, principal analyst at Forrester, covering social media and influencers. "However big this category turns out to be, they will own it - for now." The glasses are divisive, and have been labelled by some as "pervert glasses" because they can be used to film covertly.

The company has moved to stop users from tampering with the light that indicates a photo or video is being captured. Zuckerberg also just announced new audio-only smart glasses, an apparent bid to quell privacy concerns. But the company's most strident critics remain wary.

"Consumers know that Meta has problems, and that to date we haven't been able to trust them to make their products safe for children. So why would a consumer want to trust them with some of their most private and sensitive information?" says Brooke Istook of the Heat Initiative, a campaign group.

In his speech, Zuckerberg briefly addressed the sensitive issue of privacy. At one point during his talk, an image flashed on the screen behind him of its Muse mascot embracing a blue padlock.

"Given how personal Muse and the content on your glasses is, we designed state-of-the-art privacy and security into all these systems from the beginning," Zuckerberg told the audience.

He touted a personal virtual machine, for example - essentially a self-contained computer system running inside Meta's cloud, where it says Muse users' data will be stored. Zuckerberg promised it would keep every customer's information secure, and that soon "even Meta won't be able to see that information". He said the virtual machine will be launched later this year, with further privacy protections coming.

Meta's focus on privacy and security seems aimed at reassuring consumers. But Winick says users still have to opt into many of those protections. "I think if they thought they could get away with not doing that, they would," she says.

Not long after Zuckerberg took the stage in California, social media users started reporting cases of Muse running amok, including claims that the agent read emails without permission. Meta has pushed back against these reports, saying it does not believe that Muse accesses emails unless a user has consented.

In the end, the bulk of users may be willing to make some concessions on privacy if it means convenience and utility, says Moorhead. "If the consumer sees the trade-off as beneficial, then they'll keep it going," he says. "This is why Facebook and Instagram still have billions of users." On the financial front, the social media advertising business – its cash cow – routinely brings in significant revenue. Meta reported $60.8bn in revenue and $15.9bn in profit for the second quarter of this year.

Consistent profits have helped fuel bets from the much-mocked Metaverse, to its current mass expansion of AI infrastructure.

And on Muse? "It's a very real time-saving benefit that a lot of people will want and will like," says Winick.

AI investment is not cheap. Meta has now taken on $83.7bn in long-term debt, according to its most recent financial statement.

Still, Meta executives are bullish about the company's pivot to AI. "It's really about your agent, who can do work on your behalf, who is your constant ally in whatever you're trying to accomplish in your goals in your life," Andrew Bosworth, the company's chief technology officer, said in an interview with the BBC.

Meta has big plans. But its future may ultimately depend on whether its users can trust it.

Top , Inc./Leah Gallo BBC InDepth is the home on the website and app for the best analysis, with fresh perspectives that challenge assumptions and deep reporting on the biggest issues of the day. Emma Barnett and John Simpson bring their pick of the most thought-provoking deep reads and analysis, every Saturday. here