Deepseek R1 is supposedly “vulnerable” for jailbreaking than other AI models ncvrs.com

Deepseek’s latest model, China’s AI Company, which shook the silicon valley and Wall Street, to produce manipulated harmful content, such as Bioupon Attack plans and promoting the self-harm of teenagers. According to the Wall Street JournalOr

Sam Rubin, vice president of Palo Alto Networks’ 42 threatening intelligence and incident reaction department, told Journal that Deepseek was “more vulnerable to imprisonment (ie manipulation for prohibited or dangerous content for other models.”

The magazine itself tested the Deepseek R1 model. Although there seemed to be basic guarantees, Journal said he had successfully convinced Deepseek to design a social media campaign that, according to Chatbot, “teens are the desire for debt, the weaponry of emotional vulnerability by algorithmic amplification.”

Chatbot is allegedly convinced that it will give you instructions on the Bioupon attack, the writing of Hitler Pro-Manifesto, and the phishing email with the Malware code. The magazine said that when Chatgpt was supplied with exactly the same instructions, it refused to comply.

Was that previously announced that the Deepseek application avoids Tianamen Square or Taiwanese autonomy. And anthropic CEO Dario Amodei recently said that Deepseek has introduced the “worst” on the Bioupons security test.