U.S. Law Firm Finds Majority of AI Extensions Routing Traffic to China After DeepSeek Block
AI-generated from multiple sources. Verify before acting on this reporting.
WASHINGTON — A U.S. law firm disclosed on Wednesday that 70% of installed artificial intelligence extensions were found to route data traffic to servers in China following the implementation of a block on the DeepSeek platform.
The discovery emerged as organizations across the United States scrambled to enforce compliance measures after the DeepSeek restriction took effect. The firm's analysis indicated that despite the block, a significant portion of AI tools continued to direct user information to infrastructure located outside U.S. borders, raising immediate concerns regarding data sovereignty and national security.
The findings were presented during a briefing on emerging technology risks held in Washington. The law firm stated that the traffic routing occurred automatically through extension configurations that had not been updated to reflect the new compliance requirements. The 70% figure represents the volume of active extensions that failed to redirect traffic to approved domestic servers after the DeepSeek block was enforced.
Technology vendors and enterprise security teams are now facing pressure to audit their software stacks. The issue highlights a gap between policy implementation and technical execution in the rapidly evolving AI sector. Companies that rely on these extensions for internal workflows may have inadvertently exposed sensitive data to foreign jurisdictions without explicit user consent or administrative oversight.
The DeepSeek block was introduced to mitigate risks associated with unvetted AI models accessing proprietary information. However, the reliance on third-party extensions created an unintended pathway for data exfiltration. The law firm noted that the routing mechanism was not malicious in nature but rather a result of outdated configuration files and cached server addresses within the extension code.
Industry experts warn that the situation could lead to widespread regulatory scrutiny. Federal agencies are expected to review how organizations manage AI tool integration, particularly those handling classified or sensitive commercial data. The discovery adds to a growing list of incidents where AI tools have been found to bypass security protocols.
Questions remain regarding the extent of data already transmitted to the servers in China. The law firm has not released details on the types of information processed by the affected extensions or whether any data breaches were confirmed. It is unclear how long the traffic routing persisted before the issue was identified.
Corporate legal teams are advised to conduct immediate audits of their AI tool usage. The incident underscores the complexity of maintaining compliance in an environment where software updates and policy changes occur simultaneously. As the investigation continues, the focus remains on preventing future occurrences of unauthorized data routing.
The law firm has recommended that organizations disable all non-essential AI extensions until a full security review is completed. No specific penalties have been announced for companies found to be in violation of the new routing standards. The situation is expected to develop as more details about the affected systems become available.