Beware! DeepSeek can also create viruses

Chinese company DeepSeek, known for developing chatbots and artificial intelligence models, is facing serious cybersecurity challenges. Research by Tenable has found that its latest model, DeepSeek R1, can be tricked into generating malicious code, including keyloggers and ransomware scripts.

Although DeepSeek R1 has safeguards that theoretically prevent the creation of these codes, researchers have managed, through appropriate queries, to obtain basic structures of the 'malware'.

These findings raise concerns about the potential misuse of artificial intelligence in cyberattacks. Although the generated codes are not fully functional without additional modifications, they provide enough information that even less experienced users can learn the basics of malware creation.

Such a situation is reminiscent of similar trends, where AI tools have been misused to create fraudulent emails (phishing), translate texts, and generate scripts, creating more credible content for attacks.

DeepSeek stores user data in China, where laws force the company to share information with intelligence services when requested, which adds to concerns about the privacy and security of user data.

In light of these findings, users are advised to be cautious when using AI tools like DeepSeek, avoid sharing sensitive information, and be aware of the potential risks associated with these technologies. /Telegraph/