You will be able to disable AI features in Firefox 148, Mozilla has announced.
The next major update of the web browser, scheduled for release in late February, will offer an AI feature kill-switch in its new AI Controls panel.
You’ll be able to turn off Firefox’s AI features at a granular level. If you want to use some features, like on-device translations, but not others, like Google Lens image search, you can pick and choose:
If you don’t want any AI features in Firefox at all, a single ‘Block AI enhancements’ toggle acts as a kill-switch.
But notice the framing here: you’re not ‘disabling AI’, you’re ‘blocking enhancements’. This reads a little like ‘confirm shaming‘, a form of ‘deceptive pattern‘ that’ used to emotionally cue users to view their choice as self-sabotage.
Firefox AI Kill Switch Coming
“We’ve heard from many who want nothing to do with AI. We’ve also heard from others who want AI tools that are genuinely useful. Listening to our community, alongside our ongoing commitment to offer choice, led us to build AI controls,” Mozilla say.
It may seem like a surprising volte-face given that Mozilla is “rewiring” itself as an AI-first company, whereby the amount of monetised AI features it shoves in its products is the yardstick by which success is now judged.
But AI features in Firefox have not proved universally popular – a sentiment the entire “AI” hype industry is immune to acknowledging, even when the market speaks. Mozilla’s own statement tacitly acknowledges part of the reason for the backlash: a lack of purpose.
For instance, Firefox made a big deal about its AI-powered tab grouping feature. But maxing out people’s CPUs to run a local AI model to run tab content analysis in the background solely to suggest tab group names is clearly a solution in search of a problem.
—Does anyone really need “AI” to suggest names for a group of shopping tabs?
Mozilla continues to pay lip-service in filler-word heavy blog posts and expensive mini-sites about it heading a ‘rebel alliance’ of open source AI that aims to serve the public good rather than line the pockets of multi-billionaires.
But its actions? They tend to say otherwise.
It bought and subsequently killed FakeSpot, a prime example of ‘public service’ AI given it detected fake reviews on shopping products, and it continues to aggressively push its own users to use OpenAI, Anthropic and CoPilot chatbots via article summaries and sidebars.
Even so, it is great to see Mozilla relent – especially after its staff variously tried to redefine ‘opt-in’ as a ‘grey area’ and dismissed those asking for a kill-switch as ‘standing in the way of progress’ (all whilst championing ‘agency’ and ‘choice’, of course).