Zuckerberg sees WhatsApp as a model for regaining trust and outlines Meta’s strategy for AI, smart glasses, and the metaverse

WhatsApp as a benchmark for privacy

Mark Zuckerberg believes WhatsApp is the model Meta should follow to regain the trust it has lost with its users.

In an interview on Wednesday, 23, with journalist Joanna Stern at Meta Connect—the company’s annual event in Menlo Park, California—Meta’s CEO said WhatsApp is the leading private communication service and has the world’s largest end-to-end encrypted network, having pioneered bringing this technology to such a scale.

‘If you look at WhatsApp, for example, it’s the leading private communication service. It has the world’s largest end-to-end encrypted network. We were pioneers in bringing end-to-end encryption to this scale.’

According to him, the company’s new products—from smart glasses to the Muse AI agent—should follow this privacy architecture rather than replicate the model of Meta’s traditional social networking services. ‘Much of what Meta will do going forward will look like WhatsApp,’ he said, according to a summary released by Stern herself.

Muse and Meta’s cautious approach to AI

When asked whether artificial intelligence could destroy humanity, the executive replied: ‘I think if we all do a good job and act responsibly, then no. I’m quite optimistic that this will be a very positive future for everyone.’

The CEO rejected what he called the ‘destruction-filled rhetoric’ circulating in the industry. For him, the solution does not lie in a coordinated effort between AI laboratories, but in each company taking internal responsibility to act safely.

I don’t think we need some kind of industry-wide coordination to not necessarily screw this up. I think each lab needs to spend time and, when they realize there are problems, spend the time they need internally to make sure they’re proceeding safely.

Zuckerberg used the launch of Muse, a personal AI agent presented by Meta this month, as a practical example of this caution. ‘We had a version that was very good a few months ago. And we wanted to make sure we really got all the privacy and security parts right,’ he said. According to him, the delay was not a sacrifice. ‘It was the right thing for the people and the right thing for Meta.’

For him, the industry has reached a point where the technical performance of AI models is no longer the deciding factor. ‘I think we’re getting to a point where it may not matter as much how much it improves in math. It’s already very good at math,’ he said. What matters, he said, is the ability to follow instructions, understand the user’s intent, and understand human values.

The executive also disagreed with the common idea among researchers that there is a choice between making a model more capable and making it more secure. ‘When you hear people in the industry talk about there being some trade-off between capabilities and alignment, I just disagree with that. I think alignment is actually the next most important capability to get right,’ he said.

According to him, Muse was built to reflect on its own performance over time. ‘After it does something, in downtime, it looks at what happened throughout the day and can reflect on how to improve,’ he says.

Smart glasses and the company’s vision for the future

The answer came after the journalist asked about public distrust of the company’s smart glasses, which on social media were dubbed ‘creep glasses’ because they allow people to be recorded without them noticing. The controversy gained momentum last year when Meta’s product director downplayed the criticism, suggesting that ‘social norms’ would naturally form around the use of the device. During Connect, the company launched a version of glasses without a camera, in direct response to these concerns.

Apart from the issue of security, Zuckerberg reaffirmed Meta’s long-standing bet that glasses, not cell phones, will be the next major computing platform. He described the evolution as a natural sequence, from mainframes to PCs, from laptops to smartphones, always moving towards more discreet and ubiquitous devices. ‘I would say that the most important thing will end up being glasses,’ he said. ‘It’s basically an AI agent that understands you and can anticipate what you need, and can understand what’s going on in the world around you.’

According to him, the shape of glasses allows an AI assistant to ‘see what you see, hear what you hear’ and display information discreetly, without taking the user’s attention away from what is around them—the opposite of what he describes as the problem with cell phones.

For him, the same logic that transformed ordinary cell phones into smartphones should be repeated with glasses. ‘Most of the glasses people have today, you just know that at some point they will become AI glasses or smart glasses,’ he said.

Metaverse and the integration between physical and digital worlds

The executive also briefly revisited the promise that underpinned much of the company’s billion-dollar investments in recent years: the metaverse. ‘I think of the metaverse as basically the blending of the physical and digital worlds,’ he said. For him, the separation between the wealth of the digital world built over the past three decades and the physical world around people is ‘strange’ and will ‘inevitably’ disappear.