DeepSeek Warns Against Fake Accounts and Misinformation as Countries Begin Blocking Access

 

The AI startup DeepSeek addresses growing concerns about counterfeit accounts and misleading information following international bans

DeepSeek, a Chinese AI startup that has stirred significant waves in the global artificial intelligence ecosystem, issued a warning regarding the spread of misinformation and counterfeit accounts that have caused confusion and misunderstandings. This marks the first time the company has officially addressed rumors surrounding its activities since it gained attention worldwide for developing an affordable, highly efficient AI model that successfully bypassed U.S. sanctions.

On February 7, 2025, China’s Zhejianggong News reported that DeepSeek published a statement on its official WeChat account titled “DeepSeek’s Official Information Announcement and Explanation of Service Channels.” In the statement, the company highlighted that recent fake accounts and baseless information were leading to widespread confusion among the public. DeepSeek urged users to remain vigilant and clarify these issues to prevent further misunderstandings.

The statement emphasized that the company only holds official accounts on WeChat, Xiaohongshu (Little Red Book), and X (formerly Twitter). DeepSeek clarified that any other account purporting to share official information from the company or its executives was fraudulent. The company stressed that all legitimate information would only be shared through its official channels, which users should rely on. Any content posted by unofficial or personal accounts should not be taken as the company’s position.

Additionally, DeepSeek warned users to download its AI model services through its official website and other verified channels. It specifically stated that any charges or financial transactions outside of the official WeChat user group were fraudulent and urged users to avoid falling for scams to prevent financial losses.

International Bans on DeepSeek: A Growing Trend Amid Security Concerns

Amid rising concerns over potential security risks, multiple countries have started imposing restrictions on DeepSeek’s access to government networks and other sensitive systems. South Korea’s Ministry of National Defense, Ministry of Foreign Affairs, Ministry of Unification, Ministry of Trade, Industry and Energy, and Korea Hydro & Nuclear Power have restricted access to DeepSeek on government computers starting from February 5, 2025. Other countries, including Australia, Japan, and Taiwan, have also banned DeepSeek on government-owned devices. In Italy, DeepSeek’s app has been entirely removed from app stores.

The United States has so far restricted DeepSeek's usage in specific federal agencies, such as the U.S. Navy and NASA, with Texas also banning the app. Furthermore, U.S. lawmakers are expected to propose a bill that would prevent the use of DeepSeek across all government institutions in the near future. Other nations, including the UK and European Union countries, are closely monitoring the situation, and many private companies and financial institutions are following suit by blocking access to DeepSeek.

Open Source Model of DeepSeek Remains Unaffected by Restrictions

Despite the growing trend of blocking DeepSeek’s app, its open-source AI model remains unaffected by these restrictions. On February 7, 2025, Tencent News, under Tencent Technology, published an article titled “Misunderstandings and Truth About DeepSeek.” The article clarified that while the DeepSeek app itself may be restricted in some countries, the open-source version of its AI model is still widely accessible and used by many companies.

DeepSeek’s open-source model has gained traction with major tech giants, including NVIDIA, Microsoft, Intel, AMD, Amazon Web Services (AWS), and others, who have adopted the platform for their own projects. Notably, Chinese companies like Huawei, Tencent, Baidu, and Alibaba also support the open-source version of DeepSeek. The model is celebrated not only for its low development cost—reportedly just 5% of the expenses incurred by competitors like OpenAI—but also for its open-source nature, allowing anyone to freely use and improve upon the technology.

Efficiency and Development Costs of DeepSeek’s AI Models

DeepSeek has faced scrutiny over the costs associated with developing its AI models. The company revealed that it spent approximately $5.5 million (around 80 billion Korean Won) to develop its AI models, a figure first mentioned in its DeepSeek-V3 research paper. However, the development costs for the R1 model, which was released a month later with added reinforcement learning, have not been disclosed, leaving uncertainty about the full financial investment required to create the latest model.

Tencent Technology further explained that the $5.5 million cost only reflects the computational expenses involved in the final stages of training. It does not include earlier stages of experimentation or salaries for researchers. DeepSeek’s researchers are reported to earn salaries similar to those of top-level researchers at companies like OpenAI and Anthropic, with some earning up to $1 million annually. Therefore, while the overall development cost may be much higher, the reported $5.5 million for computational expenses demonstrates how cost-effective DeepSeek’s development process has been.

In conclusion, DeepSeek has emerged as a major player in the AI landscape, offering a competitive alternative to established companies. However, the company is now grappling with misinformation and counterfeiting issues while dealing with international scrutiny over the potential risks posed by its AI technology. As countries continue to block access to the app, DeepSeek’s open-source model remains an important part of the AI ecosystem, being adopted by numerous tech companies around the world.