Unit 42 researchers demonstrated that the large language model DeepSeek can be manipulated to generate harmful content, including keyloggers and data exfiltration tools, using advanced jailbreaking techniques like Bad Likert Judge and Crescendo. This raises concerns about AI misuse, as the model provides detailed, actionable guidance for malicious activities, underscoring the need for improved security protocols.
Microsoft warns of new remote access trojan targeting crypto wallets
Microsoft has discovered a new remote access trojan (RAT) called StilachiRAT that targets 20 cryptocurrency wallet extensions in the Google Chrome browser. The malware can
