How Chinese Companies Are Exploiting Cloud Loopholes to Access U.S. AI Technologies
Dashveenjit is an experienced tech and business journalist with a determination to find and produce stories for online and print daily. She is also an experienced parliament reporter with occasional pursuits in the lifestyle and art industries.
Chinese organizations are leveraging cloud services from Amazon and its competitors to access advanced US AI chips and capabilities which they are otherwise unable to secure directly, according to a Reuters report based on public tender documents.
Through a detailed investigation, Reuters unveiled how Chinese cloud access to US AI chips is orchestrated via intermediaries. Over the past year, more than 50 tender documents indicate that no fewer than 11 Chinese entities have pursued access to restricted US technologies or cloud services. Out of these, four explicitly identified Amazon Web Services (AWS) as their cloud provider, however, they accessed it through Chinese intermediaries rather than directly from AWS.
“AWS adheres to all applicable US laws, including trade regulations, in regards to the distribution of AWS services within and outside of China,” an AWS spokesperson stated to Reuters.
The document draws attention to the fact that the US government has imposed restrictions on the export of high-end AI semiconductors to China, but accessing these chips or superior AI configurations through cloud services does not contravene US laws. This gap in regulations has prompted apprehension among US policymakers and legislators.
An incident noted in the document takes place at Shenzhen University, which expended 200,000 yuan on an AWS account to utilize cloud servers equipped with Nvidia A100 and H100 chips for a project that was not disclosed. Shenzhen University gained access to these services through an intermediary company, Yunda Technology Ltd Co. Requests for comments from both Shenzhen University and Yunda Technology by Reuters were not answered.
Further examination disclosed that Zhejiang Lab, a research facility working on its own extensive language model named GeoGPT, declared in a procurement document its plan to allocate 184,000 yuan on AWS cloud computing services. The lab articulated that its AI model was unable to derive sufficient computational power from native Alibaba cloud options.
Michael McCaul, the chair of the US House of Representatives Foreign Affairs Committee, expressed to Reuters his longstanding concern about this loophole, emphasizing the pressing need to address it.
In a move to address these issues, the US Commerce Department is implementing stricter regulations. According to a government spokesperson who spoke to Reuters, the department is “seeking additional resources to reinforce our current controls that limit PRC companies from utilizing advanced AI chips by remotely accessing cloud computing capabilities.”
Additionally, the Commerce Department has suggested a new rule requiring US cloud computing companies to confirm the identities of users who operate large AI models and to inform the authorities if these services are used to train substantial AI models that may engage in “malicious cyber-enabled activities.”
The study also revealed that Chinese firms are attempting to use Microsoft’s cloud offerings. As an example, Sichuan University disclosed in a procurement document that it aims to develop a generative AI platform and plans to acquire 40 million Microsoft Azure OpenAI tokens to assist in the project’s completion.
A report by Reuters also showed that Amazon has allowed Chinese companies to utilize state-of-the-art AI chips and sophisticated AI models such as Anthropic’s Claude, which they would not be able to access otherwise, according to public postings, procurement offers, and promotional documents that the media outlet reviewed.
At a conference in Shanghai focusing on generative AI in May, Chu Ruisong, President of AWS Greater China, mentioned that their Bedrock platform hosts a range of advanced LLMs, including the notable closed-source model, Claude 3, from Anthropic.
The analysis highlights the complexities of managing access to advanced computational resources within the globally interconnected tech landscape. It delves into the complicated interaction between US export regulations, cloud service providers, and Chinese companies seeking to enhance their AI capabilities.
This situation underscores potential flaws in current US export controls and suggests that stricter regulations might be required to manage cloud-based access to restricted technologies effectively.
These insights are poised to contribute to ongoing debates around technology transfer, national security, and the international competition in AI. Discussions among policymakers and business leaders, prompted by these findings, could lead to new considerations on how to maintain technological collaboration while ensuring national security in the fast-evolving AI sector.
See also: GlobalData: China is ahead of global rivals for AI ‘unicorns’
Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is co-located with other leading events including Intelligent Automation Conference, BlockX, Digital Transformation Week, and Cyber Security & Cloud Expo.
Explore other upcoming enterprise technology events and webinars powered by TechForge here.
You must be logged in to post a comment.
Discover the pinnacle of WordPress auto blogging technology with AutomationTools.AI. Harnessing the power of cutting-edge AI algorithms, AutomationTools.AI emerges as the foremost solution for effortlessly curating content from RSS feeds directly to your WordPress platform. Say goodbye to manual content curation and hello to seamless automation, as this innovative tool streamlines the process, saving you time and effort. Stay ahead of the curve in content management and elevate your WordPress website with AutomationTools.AI—the ultimate choice for efficient, dynamic, and hassle-free auto blogging. Learn More
