DeepSeek, the Chinese artificial intelligence company, remains a hot topic in the segment. Recently, OpenAI shared a letter offering some recommendations to the US government regarding the protection of the local AI industry. In the letter, OpenAI also proposes that the US officially ban the use of DeepSeek AI in government or military segments.
DeepSeek AI and its impact on the industry
DeepSeek took the tech industry by storm by offering AI models with performance comparable to—or superior in some parameters to—that of leading AI platforms. The company also boasted that it achieved these results by investing only a fraction of what others have spent. This led to an immediate response from Wall Street in the form of stock crashes, with NVIDIA, the largest supplier of AI chips, being the hardest hit. After all, investors began to doubt whether the company’s latest hardware is really needed.
That said, experts have cast doubt on DeepSeek’s claims about development costs. Some estimate a real investment of around $1.6 billion. On the other hand, OpenAI and Microsoft have been investigating whether DeepSeek distilled data from ChatGPT to train the R1 and V3 models. This would have also helped it save on AI training costs.
OpenAI proposes US authorities ban DeepSeek from government and military offices
OpenAI had not taken an aggressive stance against DeepSeek until now. Sam Altman, the firm’s CEO, even welcomed the new competition. However, things seem to have changed quite a bit. OpenAI shared a series of proposals with the US government, including banning DeepSeek from government, military, and defense offices. The letter describes DeepSeek as “state-subsidized” and “state-controlled.”
It’s noteworthy that some US agencies and states have already banned access to DeepSeek on their own. However, there is still no executive order mandating this at the country level.
So far, there is inconclusive evidence that the Chinese government directly controls DeepSeek. However, multiple clues point to links between the government and the company. For example, the chatbot often refuses to respond to topics sensitive to Beijing. However, this may be a result of the influence of Chinese laws. Let’s remember that DeepSeek’s servers are located in China and abide by the country’s laws.
In fact, this is one of the main concerns of US AI experts. The fact that DeepSeek’s servers are located in China means that the government could request access to the platform’s user data.
Other suggestions from OpenAI
OpenAI’s proposal includes measures to protect US AI tech from potential espionage from China. The document urges the US to “ban the use of PRC-produced equipment (e.g., Huawei Ascend chips) and models that violate user privacy and create security risks such as the risk of IP theft.” The proposal contemplates that this measure also applies to the so-called “Tier 1” countries.
The company compares the risks of DeepSeek to national security with those represented by Huawei in its day. “As with Huawei, there is significant risk in building on top of DeepSeek models in critical infrastructure and other high-risk use cases given the potential that DeepSeek could be compelled by the CCP to manipulate its models to cause harm,” the letter says.
Lastly, the company also wants US authorities to allow the free use of copyrighted content for AI training. This will be necessary to maintain the country’s leadership in the sector, OpenAI claims.
2025-03-15 15:08:09