Firefox’s new AI features cause CPU spikes and drain batteries

Firefox users are experiencing unexpected CPU usage spikes and increased power consumption, with early evidence pointing to artificial intelligence features recently integrated into the popular web browser. Mozilla has acknowledged the performance issues and reversed the rollout of the problematic feature while working on a permanent fix.

The performance problems appear linked to an “inference engine” – essentially a local AI system that runs directly on users’ devices rather than in the cloud. This engine powers several new features in recent Firefox versions, most notably an AI-powered tab organization system that automatically suggests names for groups of browser tabs.

What’s causing the performance issues

Mozilla has embedded machine learning capabilities into Firefox to enhance user experience, but these AI features require significant computational resources. The most prominent addition is an AI-powered tab grouping system that analyzes open web pages to suggest relevant group names and recommend additional tabs.

The company also integrated a chatbot interface directly into the browser’s sidebar, giving users access to AI conversation tools without leaving their browsing session. Additionally, Firefox now includes enhanced address bar functionality that uses AI to help users find previously visited websites even when they can’t remember exact keywords.

These features process information locally on each user’s device rather than sending data to external servers, which Mozilla positions as a privacy advantage. However, local AI processing demands substantial CPU power and memory, particularly when running continuously in the background.

Mozilla’s response and current status

Mozilla has acknowledged the unintended performance impact and taken immediate action. “We unintentionally shipped a performance bug during the phased rollout of this feature, which processes information privately on-device,” a Firefox spokesperson explained. “After receiving reports of issues that hadn’t come up in our testing, we reversed the rollout and the performance issues should be resolved.”

The company emphasized that the problematic features were designed to protect user privacy by processing data locally rather than sending information to remote servers. However, the implementation created unexpected resource demands that weren’t apparent during internal testing.

Mozilla is currently developing a permanent fix for the performance issues while maintaining the privacy-focused approach of local AI processing.

How to disable AI features in Firefox

Users experiencing performance issues can disable these AI features through Firefox’s settings interface or advanced configuration options.

Basic method through settings interface

The simplest approach involves using Firefox’s graphical settings:

• Navigate to Firefox Settings
• Look for the gear icon (⚙️) labeled “Customize Sidebar”
• Find “AI chatbot” in the Firefox tools section
• Uncheck the box to disable the feature

Advanced configuration for complete removal

For users wanting to disable all AI-related features, Firefox provides access to advanced configuration settings:

• Type about:config in the address bar
• Search for browser.ml.chat.enabled
• Set this preference to “false”
• For comprehensive AI feature removal, search for browser.ml.enable and disable this master setting

Users can also monitor Firefox’s resource usage by typing about:processes in the address bar, which displays detailed information about each browser process and its resource consumption.

Checking for the inference process

The problematic background process appears as “Inference” in Firefox’s process list. Users experiencing unexplained CPU usage can verify whether this process is running and consuming excessive resources through the built-in process monitor.

Broader implications for browser AI integration

Firefox’s experience illustrates the challenges facing web browser developers as they integrate AI capabilities. While these features can enhance user experience, they also introduce new technical complexities and resource requirements that may not align with user expectations or hardware capabilities.

The incident highlights a broader trend across the technology industry, where companies are rapidly integrating AI features into existing products without always fully anticipating the performance implications. Mozilla’s approach of local AI processing, while privacy-focused, requires significantly more computational resources than cloud-based alternatives.

User concerns and feedback

The AI integration has generated mixed reactions from Firefox users. While some appreciate the privacy-focused approach of local processing, others have expressed frustration with both the performance impact and the perceived unnecessary nature of AI features for basic browsing tasks.

Community feedback suggests many users prefer browsers to focus on core functionality – speed, security, and privacy – rather than adding AI capabilities they didn’t request. This sentiment reflects broader user skepticism about AI integration in everyday software tools.

Looking ahead

Mozilla faces the challenge of balancing innovation with user expectations and system performance. The company’s commitment to privacy-first AI implementation sets it apart from competitors who rely on cloud-based processing, but this approach requires more sophisticated local optimization.

The performance issues serve as a reminder that AI integration requires careful consideration of computational costs, particularly for software that users expect to run efficiently in the background. As Mozilla works on fixes, the incident provides valuable lessons for other browser developers considering similar AI features.

For now, users experiencing performance issues can disable the problematic features while Mozilla develops more efficient implementations that maintain both privacy protection and system performance.