Reports of attacks that manipulate AI chatbots and pipelines have risen sharply this year.
Firms are investing in adversarial testing and robust access controls around AI infrastructure.
The findings underscore a broader shift from defending endpoints to defending AI supply chains.
