Security researchers from Cisco and the University of Pennsylvania have discovered severe safety vulnerabilities in DeepSeek’s R1 AI chatbot. According to findings published by Matt Burgess in Wired, the model failed to detect or block any of the 50 tested malicious prompts designed to elicit harmful content. The researchers achieved a 100% success rate in bypassing the system’s safety measures using standardized evaluation prompts from HarmBench. Additional testing by AI security company Adversa AI confirmed these vulnerabilities, finding that DeepSeek’s restrictions could be easily circumvented using various jailbreaking techniques. While all language models face security challenges, DeepSeek’s performance notably lagged behind competitors like OpenAI’s systems. The Chinese AI company has not responded to requests for comment about its model’s safety features.