Responsible AI auditing for LLMs and SLMs — deception, fairness, sociotechnical risk, regulatory compliance