ChainThink report: On April 9, according to a Gartner forecast report, by 2028, it is expected that 25% of enterprise-grade generative AI applications will experience at least five minor security incidents annually. The report highlights that generative AI exhibits characteristics such as poor model interpretability, high data privacy risks, and novel attack vectors, making traditional security controls insufficient to fully address emerging threats, thus presenting significant security challenges for enterprises rapidly adopting this technology.

Gartner advises enterprises to develop dedicated security strategies, including continuous monitoring of AI models, implementation of data governance frameworks, employee training on identifying AI-related risks, integration of AI security into broader cybersecurity programs, and consideration of adopting emerging AI security tools in collaboration with vendors.