Bloomberg: Pentagon employees used 'DeepSeek' for days before prohibiting

An informed military official revealed that the US Department of Defense linked their computers to Chinese servers to reach the new chat robot that was developed by Deepseek for at least two days, before the Pentagon moved to ban access to it. The official and another person who is familiar with the case asked not to mention their identity, because the information is not common, that the defense information systems agency, which is responsible for the Pentagon Information Technology Networks, occurs late Tuesday access to the Chinese Startup website. The informed person indicated that this step came after defense officials raised the fear of using the Pentagon workers of the Chinese artificial intelligence model. The Deep Seck Privacy Policy mentions that it stores user data on servers in China, and that the Chinese law rules this information. Also read: What is the Chinese “Deepseek” and why are the giants of artificial intelligence concerned? Some screens in the Pentagon showed a mark on Wednesday, saying, ‘The site is forbidden for operational reasons,’ but others can still reach ‘deeply ill’ according to the defense official, and the correspondence seen by ‘Bloomberg News’. A spokesman for the Defense Information Systems Agency did not respond to the comment. The ‘deeply ill’ earthquake has ‘deep secure’ and the artificial intelligence model caused it to develop with less than $ 6 million, an earthquake in the markets around the world earlier this week, which raised questions about hundreds of billions of dollars spent by major US technology companies on the infrastructure of artificial intelligence. Executive officials in the field of technology, including Mark Andrink, praised the Chinese model, which was first in the number of downloads in the “Apple” store. Meanwhile, security issues on the automatic chat program, including the abuse of user information stored in Chinese data centers, have escalated. Also read: Despite the investigation into his data. “Microsoft” praises the “Deepseek” model, the official said that the Pentagon IT experts still determine the nature and how the employees use the ‘deeply sick’ system directly through the browser. In the fall of 2024, the US Army staff began a previous version of “Deep Cick” on their work computers, according to the person familiar with the case. He added that these downloads at the time did not cause the concern of the security teams of the Ministry of Defense because the link in China was not clear to them. A campaign in the military sectors led interest in the latest version of “Deep Sick” to start a campaign in some military sectors, to find software instructions and to remove from individual employees’ agencies. However, thousands of employees from the Ministry of Defense Ministry “Deep Seck” used by the ‘ASK Sage’ platform for approved software, which does not directly link it to Chinese servers, according to Nicolas Shailan, founder and CEO of the platform. The military sectors deal with the use of the Chinese model by employees in different ways, as the Navy banned any use of the model on Friday due to the possible safety and moral problems associated with the origin of the model and its use, according to the US CNBC. Also read: 17 Preliminary idea of ​​the wave of stock sale due to “Deepseek”, and Navy spokesman Lauren Chattas said the Navy already leads to the use of Open Source artificial intelligence systems in official work. She added that the recent Navy correspondence, Deep Seck, called on these guidelines. The Air Force Ministry does not have specific guidelines on the model, but it already prohibits the use of sensitive general information in commercial obstetric intelligence systems without appropriate approvals, according to Air Force spokesman Laura Mc Andrewroz. In June 2024, the military issued instructions and warned about ‘unique challenges in terms of the privacy of data, safety and control of the resulting content’, according to a public memorandum, which urged the leaders to develop appropriate management operations while directly obstructing the ban on obstetric intelligence instruments. Some service staff are now discussing the possibility of issuing new policies to explicitly prohibit Chinese obstetric intelligence models, according to the correspondence seen by “Bloomberg News”.