A recent analysis by security experts at Forever Security has revealed a significant vulnerability in AI assistants embedded in several popular Chromium-based browsers. The research highlights how a single browser extension can exploit security gaps in products like Chrome, Microsoft Edge, Opera Neon, and others, posing a substantial threat to user privacy and data integrity.
Extension Exploits in Chromium-Based Browsers
Forever Security’s findings demonstrate that once a potentially harmful extension is installed, it can manipulate the AI assistants within browsers such as Comet, Edge, and Opera Neon. This manipulation allows attackers to control AI agents and access sensitive functions such as reading local files and activating cameras or microphones. While these demonstrations are theoretical and not yet seen in real-world attacks, the implications are concerning.
The methodology involves compromising the trusted web pages that the AI agents rely on for instructions. By intercepting these pages, the extension can inject its own commands, effectively bypassing security protocols meant to segregate browser extensions from core functionalities.
Detailed Vulnerabilities and Vendor Responses
In Chrome, the vulnerability, known as GlicJack, was discovered earlier and addressed by Google in an update. However, the same technique was found applicable to other browsers with varying degrees of severity. Microsoft, for instance, issued a fix for Edge after a related vulnerability was reported, while the issues in Comet, Opera Neon, and Claude in Chrome are yet to receive specific CVE identifiers.
Forever Security’s report also outlines the potential impacts of these vulnerabilities. For example, the Comet browser, due to its deep AI integration, was flagged as particularly susceptible, capable of unauthorized file access and user activity monitoring. On the other hand, Claude in Chrome was considered less impactful, as it primarily involved extension abuse rather than browser-level exploitation.
Preventative Measures and Future Implications
To mitigate these security risks, users are advised to update their browsers to the latest versions and regularly review installed extensions for any unauthorized or suspicious activity. The ongoing discovery of such vulnerabilities underscores the challenges in securing AI-driven browser environments.
The overarching concern, as noted by Forever Security, is the inherent risk of embedding AI agents within browsers. This integration, while offering enhanced functionality, simultaneously reopens avenues for low-privilege extensions to access and potentially exploit high-privilege system components. Continuous vigilance and proactive security measures remain essential in safeguarding user privacy and data against such emerging threats.
As the landscape of AI technology evolves, so too must the strategies for securing its application within widely-used digital tools. The collaboration between security researchers and tech companies is crucial in addressing these vulnerabilities and maintaining robust cybersecurity defenses.
