ai-practitioner video for a development team has deployed a generative AI model with guardrails in place. However, during testing, they discovered that cleverly
A development team has deployed a generative AI model with guardrails in place. However, during testing, they discovered that cleverly crafted inputs can still manipulate the model into producing outputs that violate the intended safety guidelines, despite restrictions. What security risk does this scenario most accurately represent?