Fighting AI with AI requires enduring, new approaches

Federal and industry experts say continuous monitoring, evaluation and red teaming can help organizations ensure their AI models are safe and secure.