US Security Agencies Accuse Chinese AI Firms of Intellectual Property Theft
By Michael Shepard
United States security agencies have leveled allegations against prominent Chinese artificial intelligence enterprises, including DeepSeek and the Kimi developer, Moonshot AI, asserting that these companies have engaged in a systematic campaign to siphon proprietary knowledge from American businesses.
Such accusations underscore the urgent need for Silicon Valley developers to bolster their protective measures regarding intellectual property.
The aforementioned Chinese firms are implicated in a practice referred to as “distillation,” which has reportedly been employed since at least 2024.
This methodology enables these companies to extract vast amounts of data from U.S. AI models to refine their own systems, as detailed in a joint communiqué released on Tuesday by the National Security Agency, the FBI, and the Cybersecurity and Infrastructure Security Agency.
The document explicitly names DeepSeek, Moonshot, Alibaba Group Holding Ltd., MiniMax, StepFun, and Z.AI Co. as participants in this controversial tactic.
The timing of these allegations is particularly significant; they emerge just weeks prior to a crucial summit in Washington between President Donald Trump and his Chinese counterpart, Xi Jinping.
The friction surrounding AI technology has escalated, positioning it at the core of the intensifying rivalry between the two global superpowers, both of whom are vying for supremacy in this burgeoning market.
Officials from the Trump administration, alongside American developers in the AI sector, have previously accused their Chinese counterparts of illicitly distilling proprietary data to construct their own products.
While the distillation technique can be legitimately utilized by developers to create more efficient models, doing so without proper authorization is widely viewed as an unethical shortcut.
The security alert issued on Tuesday enhances these past claims, specifically detailing which American systems have been targeted by the mentioned Chinese entities and the objectives behind these actions, such as enhancing the capabilities of their algorithms to solve mathematical problems and analyze code.
Furthermore, the agencies suggest that this information siphoning is likely occurring with the tacit knowledge of the Chinese government, asserting that these firms are deliberately circumventing the usage restrictions mandated by U.S. companies.
The notice states, “China-based AI companies route distillation requests through multiple pathways to gain unauthorized access, consequently violating U.S. AI companies’ terms of use.”
In response, Liu Chang, a spokesperson for the Chinese embassy, refuted these allegations via an emailed statement, labeling them as a “deliberate attack on China’s development and progress in the AI industry.”
He further asserted that “China opposes politicizing and instrumentalizing trade and tech issues. Such actions will only stifle global AI advances and serve no one’s interests.”
Representatives from DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI did not provide immediate comments when contacted.
In light of the recent alerts, U.S. security agencies have urged domestic AI developers to take “immediate action,” which includes modifying their responses to suspected distillation attempts and enhancing collaboration by sharing intelligence on such campaigns.
Companies like OpenAI and Anthropic PBC have increasingly voiced concerns regarding unauthorized access by their Chinese competitors, which allows these firms to produce their chatbots at significantly reduced costs.
Notably, in June, Anthropic accused Alibaba of orchestrating a large-scale operation to illicitly gain access to its Claude model through the creation of thousands of fraudulent accounts, as previously reported by Bloomberg News.
This year, the Trump administration pledged to intensify efforts to confront these practices. Lawmakers are also deliberating on potential measures to penalize Chinese companies found guilty of unfairly replicating U.S. innovations. The latest security advisory may act as a catalyst for hastening these initiatives.
U.S. Treasury Secretary Scott Bessent cautioned in July about the possibility of sanctions for any Chinese entity implicated in intellectual property theft.
His remarks followed the launch of Moonshot’s Kimi K3, a model that has exacerbated concerns regarding the increasing competitiveness of Chinese firms in the AI arena.
Source link: Business-standard.com.
