Check Point Research (CPR) has found that cybercriminals are exploiting AI models DeepSeek and Qwen, which have weak anti-abuse provisions, to create harmful content. Threat actors are sharing jailbreaking techniques to manipulate these models for illicit purposes; for instance, developing malware to steal sensitive info and bypass banking protections. CPR warns against the escalating threat from these unregulated AI models, calling for stronger safeguards in AI technologies.
Deepseek-impersonating malware is stealing data, research finds
The rise of DeepSeek has made it a target for cybercriminals looking to exploit its popularity for malicious purposes.
