While AI-powered networking solutions are already in widespread use by enterprises, attitudes towards AI vary and there are some fascinating contradictions in the results of our survey of 490 IT and networking professionals.
For example, respondents expect AI to make networks more vulnerable and believe a completely new security approach is needed, but they’re also confident that AI-driven defenses will outpace AI-driven attacks.
They’re worried about AI making incorrect security decisions, but they’d trust AI to autonomously quarantine the network if a confidence threshold was met.
They want to work with external partners to manage AI networking decisions and would like to outsource model training, but they want to keep responsibility for maintaining audit records entirely in-house.
These results are likely to reflect the fact that AI in networking is still comparatively new. Organizations are looking to external partners to fill skills and experience gaps that are affecting their confidence in deploying AI. But at the same time they’re worried about compliance and security risks like exposure of training data, and about the security practices and data aggregation strategies of AI providers.
Looking to the future, respondents expect cloud hyperscalers to be at the forefront of AI in networking by 2030. Predictive and simulation technologies and autonomous networking are anticipated to have a big impact on enterprise networks. Respondents don’t think things will change immediately, though, expecting that the greater part of their networks will be managed by humans traditionally or with AI assistance in the next three years.
