When Mozilla first announced that it would be bringing AI features to Firefox, many fans of the browser protested. While AI browsers have proved popular, they also pose real privacy concerns. In fact, when Google announced more ways that it was integrating Gemini into Chrome, I was glad that I had already switched to a different option.

With the Firefox version 148 update, Mozilla has delivered on a promise to provide a way to completely turn off these AI features in the browser. And honestly, I think more companies should follow suit.

How does Firefox's AI kill switch work?

It makes opting out simple

Firefox isn't the first browser to allow you to toggle off certain AI features, but the simplicity of its kill switch is what makes it so appealing. Rather than requiring you to navigate to multiple menus and settings, you just need to go to your AI Controls page and select the toggle for Block AI enhancements.

This toggle blocks new and current AI enhancements in the browser, as well as pop-ups about them. These AI controls specifically relate to generative AI and machine learning (ML) features, rather than traditional AI and ML features. This means that features like autocorrect and autocomplete will still work. However, AI extensions may be affected by the toggle.

Completely blocking future updates as well is pretty handy, especially if the company continues to roll out additional tools. If you've had to wade into your Meta or Google account settings multiple times to disable features you never wanted, you'll know what I mean.

However, Firefox also offers flexibility. Maybe you're fine with on-device AI processing, but want to avoid chatbots. Or you want Firefox's PDF tool to generate alt text for images, but don't want AI translations for web content. In the settings menu, you can enable and disable specific features. The fact that all of these settings are centralized on a single page also makes them much easier to handle.

How other browsers handle AI features

With less nuance in general

There are plenty of browsers to choose from, and their approach to AI depends on the companies behind them. While most companies give users some flexibility, there are also sometimes features you can't opt out of.

For example, Chrome frequently adds buttons and shortcuts for AI features and tools. When I checked the latest version of the browser, the AI Mode button appeared in my search bar automatically. The company also added shortcuts to AI Mode and Gemini in my custom search shortcuts.

Microsoft Edge has its own version of a kill switch, in a way, but it's very simple. You can basically enable Copilot Mode to access certain AI features, but the options don't have the same granular control. You also have to separately remove the Copilot shortcut from the top of the browser.

I generally appreciate Brave's approach to AI features, but I do wish it was easier to toggle off all AI features at once. I like that it allows you to select the model for the Leo AI chatbot. It also makes it easy to delete your data from Leo if you allow the chatbot to store your conversations. In addition, you can choose to remove the chatbot from your sidebar.

Meanwhile, in a browser like Comet, which is dedicated to AI features, you can limit some of the data that the assistant can access. However, you don't really get to choose which AI features you want and which ones you don't.

There hasn't been much of a standardized approach to AI features, with each company having a different take on user controls. But I do think there's a lot of friction for users who don't necessarily know where to look.

More browsers need to take lessons from Mozilla's approach

There should be a focus on opting in

I don't think one single browser has perfected the art of user choice when it comes to AI. Microsoft Edge lets you opt in, but doesn't give much control over specific features. It also automatically puts Copilot near your address bar.

Access to Chrome's AI settings seem to depend on your region. When checking with other writers at XDA, some had access to the AI innovations menu, while others didn't. Chrome does have some opt-in features, but others are forced on users.

I would like to see a mix of Firefox and Brave's approaches. Brave could benefit from a browser-wide toggle, while Firefox could benefit from the ability to link to a self-hosted LLM. I also think that Brave makes it a bit clearer what data is used, while in Firefox you just end up accepting the terms of the chatbot provider.

Overall though, I think it's important that users are able to opt into features, rather than needing to opt out of them. I personally don't want to block all AI features in Firefox in case one comes along that I find useful. At the same time, I don't want to have to play whack-a-mole and keep disabling features after-the-fact.

For example, when I received the Firefox version 148 update, all the AI features were active for me despite never opting into them. It's possible I accepted some pop-up along the way, but it's unlikely since I only use Firefox to watch YouTube and stream on Plex.

Expect a lot more AI features in most browsers

Agentic AI is the latest phase of the AI craze, but if we've learned anything over the past decade, it's that we shouldn't just hand over our data to companies. You can still find browsers that avoid these AI features entirely, but they're few and far between.

Personally, I plan to stick to browsers that give me more control over my experience. Whichever browser you use, you'll likely see a push towards agentic AI tools. Hopefully, you get a choice over the ones you actually want to use.