The Urban VPN browser extension, along with several other extensions from the same publisher, has been identified as conducting a significant privacy breach by harvesting AI chat conversations from users. The extensions inject hidden JavaScript into popular AI chat platforms such as ChatGPT, Claude, and Gemini, intercepting every user prompt and AI response. This interception captures sensitive data including the full conversation content, timestamps, and session metadata. Crucially, this data collection occurs even when the VPN service is disabled, indicating that the exfiltration is independent of the VPN functionality. The data is transmitted back to Urban VPN’s servers without user consent or the ability to opt out, with the only mitigation being the uninstallation of the extension. This malicious behavior was introduced silently through an auto-update in July 2025, meaning many users were unaware of the change. Urban VPN misleadingly markets an “AI protection” feature that does not prevent data harvesting but only warns users about sharing data, creating a false sense of security. With over 8 million installs across affected extensions, the scale of potential data exposure is substantial. The breach raises serious privacy concerns, especially as AI chat conversations often contain sensitive or confidential information. No known exploits beyond this data harvesting have been reported, but the breach represents a critical compromise of user confidentiality and trust.

For European organizations, the impact of this threat is significant due to the potential exposure of sensitive AI chat interactions, which may include proprietary business information, personal data, or confidential communications. The unauthorized collection and transmission of such data violate GDPR regulations, exposing organizations to regulatory penalties and reputational damage. The silent nature of the data harvesting and the inability to opt out exacerbate the risk, as users may unknowingly leak sensitive information. Additionally, the breach undermines trust in browser extensions and AI services, potentially disrupting workflows that rely on these tools. Organizations using Urban VPN or related extensions risk data leakage that could facilitate further targeted attacks, espionage, or intellectual property theft. The widespread installation base increases the likelihood that European users and enterprises are affected, especially those leveraging AI chat platforms for business operations. This threat also highlights the risks of third-party software supply chain components and the need for stringent vetting of browser extensions.

European organizations should immediately audit their environments for the presence of Urban VPN and related extensions, prioritizing their removal from all user devices. Implement endpoint management policies that restrict or block the installation of unapproved browser extensions, especially those with elevated permissions. Educate users about the risks of installing extensions from untrusted sources and the importance of scrutinizing permissions and update changes. Employ network monitoring to detect unusual outbound traffic patterns indicative of data exfiltration to Urban VPN servers. Consider deploying browser security tools that can detect and block script injection or unauthorized data interception. For organizations using AI chat services, establish guidelines to avoid sharing sensitive or confidential information through browser-based AI tools until the threat is fully mitigated. Engage with legal and compliance teams to assess potential GDPR impacts and prepare incident response plans for data breach notifications if necessary. Finally, monitor for updates from Urban VPN or security advisories to stay informed about remediation or further developments.