US-Originated Cyberattacks Target Chinese AI Model DeepSeek
Cyberattacks on China’s AI model DeepSeek, which started on January 3, have been traced back to the US, according to Chinese media reports. The attacks peaked earlier this week, with a massive brute-force assault emanating from US IP addresses, as confirmed by cybersecurity experts.
The initial phase of the attack involved Distributed Denial-of-Service (DDoS) attempts, designed to overwhelm DeepSeek’s servers and disrupt its operations. As the attacks evolved, they shifted to brute-force methods aimed at cracking user IDs and passwords to better understand the model’s functioning, a report from China’s cybersecurity firm QAX Technology Group revealed.
DeepSeek’s release of its open-source AI model, R1, earlier this month challenged the dominance of US-based giants like OpenAI, Google, and Meta. DeepSeek has gained rapid popularity, offering a language model with lower resource consumption than its Western counterparts. Its rise has contributed to significant stock losses for major Western tech firms, including Nvidia.
