Mozilla’s CEO Knows You Might Not Want AI in Firefox


When Mozilla announced that its Firefox browser would have a built-in artificial intelligence kill switch, it separated itself from the crowd of tech giants eagerly adding an onslaught of AI features to every online experience.

Mozilla CEO Anthony Enzor-DeMeo says it became apparent to him when he stepped into the role in December that the Firefox community was demanding the ability to turn off the browser’s AI features entirely.

“Our community was pretty vocal, especially during the CEO announcement, that not everyone wanted [AI],” he told me in an interview. “At its core, we want to listen to our users, and they were vocal. … It was honestly on the roadmap, but I expedited it, given the community feedback.”

But the AI kill switch, which is now available on mobile as well as desktop, has been flipped by just 1% of Firefox users to turn off AI completely. And it’s being partially used by just 3% to turn off some AI features in the browser.

Enzor-DeMeo notes that tools including AI translations have value that people want to keep. The important differentiator in Firefox is that there’s a choice, he says, pointing to Microsoft defaulting to Copilot when you search on a Windows desktop and Google installing large AI models on people’s computers without notifying them.

“I think there’s a general user sentiment of, ‘hey, I didn’t ask for that, and I didn’t choose that.’ The great thing about Firefox is … we offer choice,” he says.

Enzor-DeMeo also spoke to me about Firefox’s new Smart Window, its built-in VPN, privacy concerns in the age of AI and the browser’s fresh redesign launching in the fall with an aim of keeping the internet open and fair.

Firefox Smart Window: BYO AI

A screenshot of Firefox's Smart Window in beta

Switch from a classic browser window to an AI-powered Smart Window on Firefox in beta.

Mozilla

Smart Window, which is available in beta now, allows you to choose which AI model to use on Firefox, as well as enabling you to bring your own AI models to use in the browser.

“If you want to use ChatGPT, great. If you want to use Gemini, great. … Our sidebar allows you to use all of them,” Enzor-DeMeo said. “They all excel at different things; why do I need to be forced into one of them?” 

AI Atlas

The team at Firefox is hoping other browsers take this more AI-agnostic approach, and to use privately hosted open-source AI models, too.

Mozilla also touts the privacy of your chats in Smart Window. It says it doesn’t use any of your information to train models and automatically filters out sensitive and private data. You can then choose which data the AI model does remember about you and delete anything you don’t want it to know or turn off its memory completely.

Enzor-DeMeo noted that not only are there skeptics of AI, but that a majority of the world who cannot access it. According to some stats, around 83% of the world’s population has not used AI, and in the US, only around 3% are paying for it. He called AI “largely non-profitable,” and predicted that we’ll start seeing a lot more ads in AI services soon.

“Sometimes, especially in the tech bubble, I think we get a little bit of tunnel vision or an echo chamber of AI, AI, AI, but I think when you look at it from a global scale, there’s not a ton of access,” he told me. “If we actually go the route that AI becomes more centered in the browser, and that’s how people access the internet, you run the risk of the internet becoming more closed off.”

AI tools need a lot of information about you to provide an optimal experience — “that’s just the brutal truth.” But he said Firefox focuses on active consent.

“I think there’s an inflection point in the market with lack of trust in Big Tech,” Enzor-DeMeo said. “There’s an inflection point in what is AI doing to society, and I think people crave control, autonomy, choice and essentially, privacy.”

Firefox’s built-in VPN: 1.5 million signups

A screenshot of Firefox's built-in VPN

Firefox has a built-in VPN.

Mozilla

To help maintain this privacy, Firefox introduced a free in-browser VPN last month. Enzor-DeMeo said one of his top priorities when he took over as chief executive in December was to create a built-in VPN product, because it’s a lot easier for people to just click a button in the browser than it is to open another app and log in — “a subpar experience,” he noted.

While many VPNs offer browser extensions to simplify the process, keep in mind that a browser-based VPN generally only encrypts your activity within that browser, not in other apps across your device. You need VPN apps for more robust privacy protections.

Firefox’s VPN now has 1.5 million signups, with Mozilla currently offering an Unlimited VPN package from June 9 through Aug. 31 that lets you select your geolocation. And with around 800,000 active users already, Enzor-DeMeo said it’s a nod of approval that Firefox did the right thing, especially in the age of AI running rampant, when the privacy enabled by using a VPN is incredibly important.

“What we’re seeing is an increase in surveillance. There’s a lot of different challenges, depending on your geolocation, going on with AI,” he told me. “We’ve always been a firm believer in VPN. I think people have a right to privacy. I think that people need to look up information on medical issues or things like that, and remain hyper private. So for me, it was ethos-led.”

The new Firefox: “Keep the internet an equal playing field”

A screenshot of the Firefox redesign, called Project Nova internally

The Firefox redesign has vertical tabs and rounded icons.

Mozilla

Project Nova, which will simply be called Firefox when it launches later this year, is centered around making Mozilla’s browser speedier (page load times are up to 9% faster than they were previously, Mozilla says), more secure and customizable. Firefox has around 200 million monthly users. (It’s estimated that Firefox has just over 2% of the browser market share, compared to Google Chrome’s 70% and Apple’s Safari’s 16%.) 

It’s slated to release in the fall, around September or October, although new features are slowly rolling out in Nightly. Mozilla has been seeking input from its community of developers and other users via its own forums and AMA posts on Reddit.

One new AI feature is tab groups, in which artificial intelligence will automatically group together similar tabs you have open to make them easier to find. Enzor-DeMeo said his AI approach is to find efficiency features that people would actually benefit from.

Beyond AI, the new Firefox will get visual upgrades including compact mode; a round shape for panels, menus, settings and browser controls; a glow around your active tab; and accessibility features built in.

“What we’re trying to do is keep the web open, safe and competition fair,” Enzor-DeMeo said. “Our objective is not to be the biggest browser; it’s to keep the internet an equal playing field.”





Source link

Leave a Reply

Subscribe to Our Newsletter

Get our latest articles delivered straight to your inbox. No spam, we promise.

Recent Reviews


Another day, another politically motivated attack in the United States.

This morning’s shooting at a Dallas ICE detention facility – where a sniper killed two detainees and wounded another before taking his own life prompted me to revisit a question that’s been troubling me: Is political violence actually increasing in America, or does it just feel that way?

To explore this, I’ve conducted what I’ll call a methodological experiment.

Rather than relying on traditional datasets, I’ve used ChatGPT and Claude to construct a synthetic index of political violence in the US since 1945. Let me be absolutely clear: this isn’t conventional data. It’s data generated through language models, with all the limitations that implies.

The Methodology (and Its Limitations)

Here’s what I did: I asked both ChatGPT and Claude to generate lists of politically motivated violent incidents since 1945, then had them score each incident’s severity on a scale where 50 represents a “normal” level.

The models assessed both casualties and symbolic significance, and I used them to cross-check each other’s work. I then quality-checked the output myself and categorised perpetrators by political affiliation where this was clearly established.

This approach is, admittedly, unorthodox. Language models are trained on existing texts and may reflect biases in their training data. They might overweight highly publicised events or recent incidents that featured prominently in their training corpus.

The “data” we’re looking at is essentially a structured synthesis of what these models have absorbed about American political violence.

Yet there’s something intriguing here. These models have processed vast amounts of information about political violence – news reports, academic studies, government documents. Their output might capture patterns that traditional datasets miss, though it might also amplify certain narratives or blind spots.

What the Synthetic Data Reveal

With those caveats firmly in mind, the patterns that emerge from this exercise are concerning. The model-generated index shows a clear upward trend in political violence over the past decade.

Looking at the breakdown by perpetrator ideology (where clearly established), the data suggest that right-wing extremist groups have been responsible for the majority of incidents in recent years, though we cannot draw conclusions about today’s attack whilst investigations are ongoing.

The synthetic data align with some empirical observations. Princeton’s Bridging Divides Initiative recorded over 600 incidents of threats and harassment against local officials in 2024 – a 74% increase from 2022. The University of Maryland found that in the first half of 2025, 35% of violent events targeted U.S. government personnel or facilities – more than twice the rate in 2024.

The Charlie Kirk Assassination and Recent Patterns

The September assassination of conservative activist Charlie Kirk marked a particularly dark moment.

The incident followed numerous recent acts of political violence, including the murder of Minnesota Democratic state Rep. Melissa Hortman and her husband, and two assassination attempts on President Trump in 2024.

What the synthetic data reveal is not just increased frequency but a shift in patterns. While overall levels of physical political violence remained low in 2024 compared to years prior, acts of vigilante violence grew as a proportion of all reported incidents.

We’re seeing less organised group violence and more lone-wolf attacks – a pattern that’s harder to predict and prevent.

The Epistemological Challenge

When we use language models to generate “data” about social phenomena, what exactly are we measuring? We’re essentially extracting structured information from the collective corpus of human writing about these events. It’s aggregating distributed information, but through an AI intermediary rather than traditional data collection methods.

This raises fascinating questions.

The models suggest that right-wing extremist violence has been responsible for a fairly large majority of U.S. domestic terrorism deaths since 2001. But how much of this reflects actual patterns versus the way these events are covered and discussed in the sources the models were trained on?

The synthetic data are, in a sense, a mirror of our collective discourse about political violence. They reflect not just what happened, but how we’ve talked about what happened. That’s both a limitation and, potentially, a feature – understanding the narrative landscape around political violence might be as important as counting incidents.

An Experimental Tool

I’ve built an interactive app (using the AI coding tool Lovable) based on this language model-generated violence index.

Users can explore the synthetic data, examine patterns across different time periods and perpetrator groups, and understand the methodology behind it. Think of it as an experiment in using AI to structure historical information rather than a definitive dataset.

The value isn’t in treating this as gospel truth, but in what it reveals about how these events are recorded, remembered, and synthesised in our collective digital memory.

When language models trained on our civilisation’s text output show rising political violence, it tells us something – even if that something is as much about narrative as about underlying reality.

This morning’s tragedy in Dallas reminds us that behind every data point – whether traditionally collected or AI-generated – there are real victims and real consequences. Understanding the patterns, however imperfectly, is the first step toward addressing them.

Try the tool here.





Source link