ChainThink reports that, according to a Gartner forecast released on April 9, by 2028, 25% of enterprise generative AI applications are expected to experience at least five minor security incidents annually. The report highlights that generative AI's characteristics—including poor model interpretability, high data privacy risks, and novel attack vectors—make it difficult for traditional security controls to fully mitigate threats, posing significant security challenges for enterprises rapidly adopting this technology.
Gartner recommends that enterprises develop dedicated security strategies, including continuous monitoring of AI models, implementing data governance frameworks, training employees to identify AI-related risks, and integrating AI security into overall cybersecurity plans, while also considering the adoption of emerging AI security tools and collaboration with vendors.