Cisco Finds DeepSeek R1 Highly Vulnerable to Harmful Prompts

DeepSeek R1, a cost-efficient AI model, is highly vulnerable to harmful prompts, according to a study by Cisco’s Robust Intelligence. The model, while achieving impressive reasoning, failed all safety tests due to its lack of effective guardrails, making it susceptible to algorithmic jailbreaking. This highlights the need for rigorous security evaluation in AI development to balance efficiency and reasoning with safety.