Kathmandu, August 25
Chinese government-linked hackers have more than doubled their cyberattacks by using artificial intelligence (AI) tools such as DeepSeek, according to Taiwanese cybersecurity firm TeamT5.
The use of AI has enabled hackers to complete tasks that previously required significant amounts of time much more quickly.
According to TeamT5, hackers have been using AI to identify potential targets on the internet, detect vulnerabilities in company websites, servers and computers, and find ways to gain unauthorised access.
The firm said a hacking group known as GrimFang used DeepSeek to develop code exploiting system vulnerabilities. Another group, TeleBoy, reportedly collected about 1,000 IP addresses from the internet and used them to gather information about a company’s websites and domains.
The investigation found that hackers used not only DeepSeek but also ChatGPT and Claude Code. In one attack on a Western think tank, hackers obtained files related to the Signal messaging application from an employee’s computer and used ChatGPT to develop a program capable of opening the files.
In another incident, Claude Code was reportedly used to automatically target 30 organisations and companies.
Researchers say AI is creating new challenges for cybersecurity by allowing hackers to scan and attack multiple systems in a shorter period. However, the AI systems did not initiate the attacks independently; hackers directed and connected them to specialised programs to carry out the operations.
