How to Turn Off AI in Firefox (And Why You Might Want To)
Mozilla just announced a master switch to disable Firefox's AI features. Here's how to use it — and why controlling AI in your software matters.
Mozilla announced yesterday that the next Firefox release will include a single switch to disable all AI features. No digging through about:config. No hunting through multiple settings pages. One toggle to turn it all off.
This is how it should work everywhere.
What AI Features Are We Talking About?
Firefox has been gradually adding AI capabilities over the past year:
AI-powered suggestions — Firefox suggests search completions and content based on your browsing patterns, powered by machine learning models.
Tab grouping assistance — AI that suggests how to organise your tabs based on content similarity.
Alt text generation — Automatically generates descriptions for images, useful for accessibility.
Translation improvements — Neural machine translation for pages in other languages.
Sidebar AI chat — Integration with various AI assistants directly in the browser sidebar.
Some of these are useful. Some feel like features looking for a problem. And some people simply don't want AI touching their browsing experience at all.
Until now, disabling these meant finding each feature's individual toggle — scattered across different settings pages — or diving into about:config to flip obscure flags.
The New Approach
Starting with the next Firefox release (expected mid-February), there will be two options:
Disable all AI features — One switch that turns everything off. The AI functionality simply doesn't run.
Manage individually — Keep the granular controls for people who want, say, translation but not tab suggestions.
Mozilla says they added this after user feedback. People wanted a clear way to opt out entirely without playing whack-a-mole with individual features.
How to Use It (Once It's Available)
When the update lands:
- Open Firefox Settings (
Cmd + ,on Mac, or Menu → Settings) - Look for the new "AI Features" or "Firefox Labs" section
- Toggle "Enable AI features" off
That's it. Firefox returns to a more traditional browsing experience.
For now, if you want to disable current AI features, you'll need to do it piecemeal:
Sidebar AI: Settings → Firefox Labs → Uncheck "AI chatbot"
Search suggestions: Settings → Search → Uncheck "Provide search suggestions"
Tab suggestions: Settings → General → Tabs → Uncheck related options
Translation: Settings → General → Language → Manage translation preferences
Why Would You Turn Off AI?
A few reasons, depending on what matters to you:
Privacy Concerns
AI features often need to send data somewhere to work. Firefox's translation, for example, can work locally for some languages but needs server-side processing for others. Tab grouping analyses your browsing patterns.
Mozilla is generally trustworthy on privacy — certainly more than Google — but "don't process my data at all" is the only guarantee.
Resource Usage
AI models consume memory and CPU cycles. On older hardware or when running many applications, turning off AI features can improve performance.
Simplicity
Some people just want a browser that loads web pages. Every additional feature is complexity, potential bugs, and interface clutter. The minimalist approach has real value.
Philosophical Objections
Maybe you're concerned about AI's environmental footprint. Maybe you object to how training data was collected. Maybe you just don't want machine learning algorithms mediating your relationship with the web.
Whatever the reason, you shouldn't need to justify it. "I don't want this" is enough.
The Broader Pattern
Firefox's move highlights something important: we need better controls over AI features in all our software.
Right now, AI is being added to everything. Operating systems, office suites, email clients, note-taking apps, design tools, developer environments. Often these features appear in updates without fanfare, enabled by default.
This creates problems:
Consent issues — You agreed to use a text editor. You didn't necessarily agree to have your writing analysed by AI models.
Data flow uncertainty — When does AI processing happen locally? When does data go to servers? The answers are often unclear, buried in privacy policies nobody reads.
Feature creep — Your simple tool becomes complicated. Interface changes to accommodate AI. The thing you liked about the software gets subsumed by features you didn't ask for.
Dependency — Once you rely on AI features, you're locked into them. Turn them off later and your workflow breaks.
How to Handle AI Features Generally
A framework I use when software adds AI:
1. Understand What's Happening
Before enabling (or keeping enabled) any AI feature, figure out:
- Does it process data locally or in the cloud?
- What data does it use?
- Can you use the software fully without it?
2. Start Disabled
When possible, turn AI features off initially. Use the software without them. Then enable specific features if you actually want them.
This is the opposite of the default, where everything's on and you disable later. But it gives you control over what AI touches.
3. Check Privacy Settings After Updates
Major updates often enable new features by default. After updating any software, spend two minutes checking settings. It's annoying, but necessary.
4. Prefer Local Processing
When you do use AI features, prefer ones that run on your device. Local processing means your data stays with you. Cloud processing means trusting someone else.
5. Keep Alternatives Available
If you're using AI-assisted features, make sure you can still do the task without them. Don't become dependent on something you can't control.
Software That Does This Well
Firefox isn't alone in offering proper AI controls. Some positive examples:
Apple Intelligence — Apple's AI features require explicit opt-in and clearly distinguish local versus server processing. You can disable specific features or all of them.
Obsidian — AI features are plugins you install separately. Don't install them, and they don't exist. Clean separation.
Signal — No AI features to worry about. They've explicitly stated they're not adding generative AI, prioritising security and simplicity.
Software That Does This Poorly
Microsoft Office/365 — Copilot is heavily pushed, settings are scattered across multiple locations, and it's not always clear what's using AI and what isn't. Enterprise admins get better controls than individual users.
Google Workspace — Similar issues. AI features appear throughout, with controls that are neither obvious nor comprehensive.
Adobe Creative Cloud — Generative AI is now deeply integrated. You can disable some features, but the line between "AI-assisted" and "traditional" tools is blurry.
GDPR and AI Features
For those of us in Europe, GDPR provides some leverage here.
Under GDPR, automated decision-making has specific rules. If AI features are making decisions that affect you (even something like "which tabs to suggest grouping"), you have rights around:
- Being informed that automated processing is happening
- Requesting human intervention
- Objecting to purely automated decisions
In practice, browsers don't make "decisions" in the GDPR sense — they're tools, not decision-makers. But the spirit of the regulation supports the idea that you should control when AI processes your data.
More directly: GDPR requires consent for non-essential data processing. AI features that send data to servers should require explicit opt-in, not opt-out.
If you're uncomfortable with AI features and the software makes it hard to disable them, consider that a compliance concern worth raising.
My Setup
For what it's worth, here's how I handle AI features in my daily tools:
Browser (Firefox): Will use the new toggle once available. Currently, most AI features disabled manually except translation (useful for work).
Writing (Obsidian): No AI plugins installed. I want my notes to be mine.
Email: AI drafting disabled. I write my own emails, thanks.
Development tools: Copilot enabled in VS Code — this is where AI actually helps me significantly. But I review every suggestion before accepting.
Operating system: Apple Intelligence partially enabled — on-device features like improved Siri are fine, but I've disabled anything requiring server processing for now.
The pattern: enable AI where it provides clear value, disable it where it's just overhead or where privacy concerns outweigh benefits.
The Right Attitude
AI isn't inherently bad. It's a tool. But like any tool, you should control when and how you use it — not have it imposed on you.
Mozilla's approach with Firefox is a good model: make it easy to enable, easy to disable, and respect whatever choice the user makes. No dark patterns, no buried settings, no re-enabling on updates.
More software vendors should do this. Until they do, stay vigilant. Check settings after updates. Assume new features are enabled by default. Take the time to configure things how you want them.
Your software should work for you, not experiment on you.
When the Firefox update drops, I'll update this post with exact steps and screenshots. For now, keep an eye on your Firefox settings, and consider whether the AI features you're currently using are ones you actually chose.