U.S. Navy Bans Sailors From Using Chinese AI Model Over Security Concerns

The U.S. Navy has prohibited its personnel from using the AI model developed by the Chinese company DeepSeek due to security and ethical concerns. In an email sent to sailors on Friday, the Navy warned that DeepSeek’s AI should not be used “in any capacity” for work-related tasks or personal activities.

A Navy spokesperson confirmed the email’s authenticity, stating that the restriction was based on the Department of the Navy’s generative AI policy.

The Navy’s Operational Command issued the directive as a precaution against the risks linked to AI models developed in China. The message, sent to the distribution list “OpNav” (Operational Navy), advised sailors to “refrain from downloading, installing, or using the DeepSeek model.”

Commander Tim Hawkins, a Navy spokesman, later clarified that the email was meant as a general reminder that sailors are prohibited from using any publicly accessible, open-source AI models unless approved. He stated that DeepSeek was mentioned as the most recent example of how this policy applies.

The directive follows a September 2023 Navy memo that had already warned against the use of generative AI unless its security risks were fully assessed. The memo stated that AI models save all prompts, posing a risk of inadvertently exposing sensitive or classified information.

The Office of the Director of National Intelligence’s February 5 threat assessment described China as the “most active and persistent cyber threat” to U.S. government and private-sector networks.

DeepSeek, a startup based in Hangzhou, China, launched its AI Assistant app earlier this month, which quickly became the most downloaded free app on the U.S. Apple App Store, surpassing OpenAI’s ChatGPT.

The AI model, DeepSeek R1, is open-source and reportedly performs as well as its more expensive competitors despite being much cheaper.

The company revealed that it developed its large language model in just two months with a budget of under $6 million. This is a fraction of the cost incurred by U.S. AI firms like OpenAI, Google, and Anthropic.

The model’s efficiency has raised concerns in the U.S. tech industry, with analysts warning that future AI models could require far less expensive infrastructure than previously assumed.

The success of DeepSeek’s AI shook the U.S. stock market. On Monday, shares of AI chipmakers NVIDIA and Broadcom dropped 17%, resulting in a combined market loss of $800 billion.

The Nasdaq also fell by 3.1%, driven by concerns that lower-cost AI models could disrupt existing market leaders.

Industry figures have responded to DeepSeek’s rapid rise. OpenAI CEO Sam Altman acknowledged that the Chinese model was impressive, particularly given its lower cost, but said that OpenAI is committed to delivering superior AI technology.

Meanwhile, Meta reportedly launched four “war rooms” within its generative AI division to address the challenge posed by DeepSeek.

Alexandr Wang, CEO of Scale AI, described DeepSeek’s latest model as “earth-shattering” and stated that it is “roughly on par with the best American models.” He also referred to the U.S.-China AI rivalry as an “AI war.”

U.S. President Donald Trump called DeepSeek’s rise a “wake-up call” for American tech companies. Trump’s administration recently announced an initiative called Stargate, a joint venture between OpenAI, Oracle, and SoftBank, aimed at investing billions in AI infrastructure within the U.S.

David Sacks, Trump’s AI and crypto advisor, also commented on the issue, stating that the AI race will be highly competitive and that the U.S. cannot afford to be complacent.

Meanwhile, DeepSeek announced on Monday that it was temporarily limiting user registrations due to “large-scale malicious attacks” on its services. However, the company later resumed normal operations.

After CNBC reported on the Navy’s ban, a Navy spokesperson clarified that the email had been misinterpreted as a specific ban on DeepSeek. The Navy explained that the message was intended to remind sailors of existing policies prohibiting the use of unapproved AI tools.

Despite this clarification, concerns remain about the security risks posed by AI technology developed in China. The U.S. government is closely monitoring and assessing possible threats from foreign AI models.

References: FirstPost, CNBC

Source: Maritime Shipping News