US State Department Initiates Global Campaign to Address Chinese IP Theft Allegations
The United States State Department has launched an extensive global initiative to spotlight what it characterizes as pervasive endeavors by Chinese enterprises, notably AI startup DeepSeek, to appropriate intellectual property from American artificial intelligence laboratories. This development stems from a diplomatic cable reviewed by Reuters.
Dated Friday, the cable dispatched to US diplomatic and consular posts worldwide directs personnel to engage in discussions with foreign officials regarding “concerns over adversaries’ extraction and distillation of US AI models.”
“A separate diplomatic request and communication have been conveyed to Beijing for high-level discourse regarding China,” the document elucidates.
Distillation refers to the technique of training compact AI models through the output derived from larger, more resource-intensive models, aimed at reducing the costs associated with developing powerful new AI tools.
This week, the White House echoed similar allegations. The State Department has yet to provide an immediate response regarding inquiries for additional commentary.
OpenAI has cautioned US lawmakers that DeepSeek is targeting the ChatGPT creator and prominent American AI firms to replicate models for its proprietary training purposes.
China Denies Allegations
The Chinese Embassy in Washington, D.C., reiterated its position on Friday, maintaining that the accusations lack merit.
“The claims suggesting that Chinese entities are pilfering American AI intellectual property are unfounded and represent calculated attacks on China’s advancements in the AI sector,” it stated in a communiqué to Reuters.
DeepSeek, whose affordable AI model astonished the global community last year, launched a preview of its eagerly awaited new V4 model on Friday, tailored for Huawei chip technology, underscoring China’s growing independence in this domain.
DeepSeek has not yet responded to inquiries for comment. The company has previously maintained that its V3 model utilized naturally occurring data obtained via web crawling, asserting it did not intentionally employ synthetic data generated by OpenAI.
Numerous Western and certain Asian nations have prohibited their institutions and officials from utilizing DeepSeek’s technology, citing concerns regarding data privacy. Despite this, DeepSeek’s models remain among the most prevalent on international platforms hosting open-source models.
The State Department’s cable aims to “caution against the perils of utilizing AI models distilled from US proprietary systems, and to establish the groundwork for potential follow-up actions by the US government.”
The cable also referenced Chinese AI companies Moonshot AI and MiniMax, neither of which has responded to requests for comment.
The document articulated, “AI models developed through covert, unauthorized distillation efforts empower foreign entities to introduce products that may exhibit comparable performance on select benchmarks at significantly reduced costs, yet fail to replicate the original system’s full efficacy.”
Moreover, it indicated that these campaigns “systematically dismantle security protocols embedded within the resultant models and negate safeguards designed to ensure that those AI systems strive for ideological neutrality and accuracy.”
The White House’s accusations, along with the diplomatic cable, emerge just weeks prior to US President Donald Trump’s anticipated meeting with Chinese President Xi Jinping in Beijing.
This situation threatens to exacerbate tensions in an ongoing technological rivalry between the two superpowers, a dynamic that had recently shown signs of amelioration following a détente brokered in October.
Source link: Dawn.com.
