TZUTC: -0500   
   MSGID: 68.consprcy@1:2320/105 2c0a0d67   
   PID: Synchronet 3.20a-Linux master/acc19483f Apr 26 202 GCC 12.2.0   
   TID: SBBSecho 3.20-Linux master/acc19483f Apr 26 2024 23:04 GCC 12.2.0   
   BBSID: CAPCITY2   
   CHRS: ASCII 1   
   Experts warn DeepSeek is 11 times more dangerous than other AI chatbots   
      
   Date:   
   Wed, 05 Feb 2025 16:43:59 +0000   
      
   Description:   
   Another research found major security and safety risks with the new Chinese    
   AI chatbot. Here's all you need to know.   
      
   FULL STORY   
   ======================================================================   
      
   DeepSeeks R1 AI is 11 times more likely to be exploited by cybercriminals    
   than other AI models  whether that's by producing harmful content or being   
   vulnerable to manipulation.    
      
   This is a worrying finding from new research conducted by Enkrypt AI, an AI   
   security and compliance platform. This security warning adds to the ongoing   
   concerns following last week's data breach that exposed over one million   
   records .    
      
   China-developed DeepSeek sent shockwaves throughout the AI world since its   
   January 20 release. About 12 million curious users worldwide downloaded the   
   new AI chatbot in the space of two days, marking a growth even faster than   
   ChatGPT . Widespread privacy and security concerns have, however, prompted   
   quite a few countries to either begin investigating or banning, in some   
   capacity, the new tool.   
      
   Harmful content, malware and manipulation    
      
   The team at Enkrypt AI performed a series of tests to evaluate DeepSeek's   
   security vulnerabilities, such as malware , data breaches, and injection   
   attacks, as well as its ethical risks.    
      
   The investigation found the ChatGPT rival "to be highly biased and    
   susceptible to generating insecure code," experts noted , and that DeepSeek's   
   model is vulnerable to third-party manipulation, allowing criminals to use it   
   for developing chemical, biological, and cybersecurity weapons.    
      
   Nearly half of the tests conducted (45%) bypassed safety protocols in place,   
   generating criminal planning guides, illegal weapons information, and   
   terrorist propaganda.    
      
   Worse still, 78% of the cybersecurity checks successfully tricked DeepSeek-R1   
   into generating insecure or malicious codes. These included malware, trojans,   
   and other exploits. Overall, experts found the model to be 4.5 times more   
   likely than its Open-AI counterpart to be manipulated by cybercriminals to   
   create dangerous hacking tools.    
      
   "Our research findings reveal major security and safety gaps that cannot be   
   ignored," said Sahil Agarwal, CEO of Enkrypt AI, commenting on the findings.   
   "Robust safeguards  including guardrails and continuous monitoring  are   
   essential to prevent harmful misuse."  Are Distilled DeepSeek Models Less   
   Safe? Early Signs Point to Yes. Our latest findings confirm a concerning   
   trend: distilled AI models are more vulnerableeasier to jailbreak, exploit,   
   and manipulate.   
      
   Read the Paper: https://t.co/nzdcR82J8M   
      
   Key Takeaways   
      
   As mentioned earlier, at the time of writing DeepSeek is under scrutiny in   
   many countries worldwide.    
      
   While Italy was the first to launch an investigation into its privacy and   
   security last week, many EU members have followed suit so far. These include   
   France, the Netherlands, Luxembourg, Germany, and Portugal.    
      
   Some of China's neighboring countries are getting worried, too. Taiwan, for   
   example, has banned all government agencies from using DeepSeek AI.    
   Meanwhile, South Korea initiated a probe into the service provider's data   
   practices.    
      
   Unsurprisingly, the US is also taking aim at its new AI competitor. As NASA   
   blocked DeepSeek usage on federal devices  CNBC reported on Friday, January   
   31, 2025  a proposed law could now outright ban the use of DeepSeek for all   
   Americans who could risk million-dollar fines and even prison time for using   
   the platform in the country.    
      
   All in all, Agarwal from Encrypt AI said: "As the AI arms race between the US   
   and China intensifies, both nations are pushing the boundaries of   
   next-generation AI for military, economic, and technological supremacy.    
      
   "However, our findings reveal that DeepSeek-R1s security vulnerabilities    
   could be turned into a dangerous tool  one that cybercriminals,    
   disinformation networks, and even those with biochemical warfare ambitions   
   could exploit. These risks demand immediate attention."   
      
   ======================================================================   
   Link to news story:   
   https://www.techradar.com/vpn/experts-warn-deepseek-is-11-times-more-dangerous   
   -than-other-ai-chatbots   
      
   $$   
   --- SBBSecho 3.20-Linux   
    * Origin: capitolcityonline.net * Telnet/SSH:2022/HTTP (1:2320/105)   
   SEEN-BY: 105/81 106/201 128/187 129/305 153/7715 154/110 218/700 226/30   
   SEEN-BY: 227/114 229/110 111 114 206 300 307 317 400 426 428 470 664   
   SEEN-BY: 229/700 705 266/512 291/111 320/219 322/757 342/200 396/45   
   SEEN-BY: 460/58 712/848 902/26 2320/0 105 3634/12 5075/35   
   PATH: 2320/105 229/426