DEEPEEEK's cyberatta was performed by hackers from the USA, Chinese media say
According to the CCTV State TV and Radio Company, the attack began on January 3 and reached a peak on Monday and Tuesday, while the main attempts were made from IP addresses in the United States. Cyberatta was directed to Deepeseek, which recently launched a model of AI assistant, designed to work with lower costs compared to available analogues in the US. Cyberataka has grown rapidly, causing concerns about the safety of Deeceek operations and a wider impact on the growing AI industry in China.
DEEPSEEK attack took place in several stages. In an early stage, the company faced the "distributed refusal" attaches (DDOS), as reported by South China Morning Post. These attacks work, overloading the system with excessive Internet traffic, which complicates the normal functioning of the servers. According to experts with XLab (through Global Times), Chinese cybersecurity company, cyberattacks have intensified both in scale and complexity.
DDOS-Athamies were the initial phase, but soon they gave way to attempts. These selection attacks were directed at the evil of users' identifiers and passwords, which potentially enabled the attackers to access the Deeseek platform and to understand its basic artificial intelligence technology. Attacking the password by selecting a password involves systematic testing of all possible combinations of passwords until it is found correct.
Having access to user accounts, the attacker can make himself for legitimate users and get an idea of how the system works. Van Huai, a QAX Technology Group Cybersecurity Expert, said CCTV: "All attacks have been registered, all of them from the United States. " This statement emphasizes the geographical origin of the attack that was traced to the United States.
XLab reports indicate that the source of these attacks was widespread, with IP addresses from the US, Singapore, the Netherlands, Germany and even within China. These conclusions draw a picture of a complex and well -coordinated attack on the growing Deeseek platform. In response to Cyberatics, Deepeseek decided to limit new registration for users with mobile numbers of mainland China. This step was necessary because of significant problems with registration caused by prolonged attacks.
Cyberaratak coincided with the launch of the open source code of AI-assistant DEESEEEK, which has caused a stir in the world community of experts from AI. The model, released on January 10, was highly appreciated for its ability to provide productivity similar to the performance of the most advanced models with a closed source code, but for a small part of the cost. This development has become an important milestone in the AI race and attracted the attention of US politicians.