DeepSeek’s rapid success combined with its claim that its AI model was trained for only $5.6 million sent shockwaves through the tech industry. (Alex Wilson/Stars and Stripes)
The U.S. Navy says internal correspondence about the use of generative AI has been misinterpreted as a specific ban against DeepSeek AI Assistant, a Chinese-developed program that recently sent shockwaves through the tech industry.
CNBC reported Tuesday that an email sent to sailors on Friday directed them to avoid using the app “in any capacity” due to “potential security and ethical concerns,” since the AI model originates in China.
An unidentified Navy spokesperson confirmed the email’s authenticity, CNBC said, and referenced an earlier Department of Navy memorandum on the use of generative AI.
However, Friday’s correspondence was intended to remind sailors that they’re prohibited from using any “publicly accessible, open-source AI programs or systems,” Cmdr. Tim Hawkins, a Navy spokesman, told Stars and Stripes by email Wednesday.
“DeepSeek was mentioned [in the correspondence] as the most recent example of how the standing guidance applies,” he wrote.
A memo issued in September 2023 cautions against the use of AI in an operational setting unless security requirements “have been fully investigated, identified and approved for use within controlled environments.”
It further warns that AI learning models save all prompts and that the use of proprietary or sensitive information “poses a unique security risk” that “could lead to an inadvertent release of sensitive or classified information.”
The most recent correspondence was meant to remind sailors about this memo, Hawkins said.
The U.S. government views China as the “most active and persistent cyber threat to U.S. Government, private-sector, and critical infrastructure networks,” according to the Office of the Director of National Intelligence’s annual threat assessment Feb. 5.
“If Beijing believed that a major conflict with the United States were imminent, it would consider aggressive cyber operations against U.S. critical infrastructure and military assets,” the report said.
DeepSeek released its AI Assistant on Jan. 10 and quickly outpaced OpenAI’s ChatGPT by becoming the No. 1 free app in Google Play and Apple’s App Store. It had more than 5 million downloads from Google Play as of Wednesday.
DeepSeek’s rapid success combined with a claim that its AI model was trained for only $5.6 million – a fraction of the billions spent training ChatGPT and other AI models – sent shockwaves through the tech industry.
U.S. tech stocks were also affected. Heavily impacted was Nvidia, a major producer of AI hardware, which saw share prices plummet approximately 17% on Monday - a $600 billion market value loss – BBC reported Wednesday.
The stock appeared to stabilize by Monday, rising 8.8%, the report said.
