The fact that they are already naming it an AI species shows an intent.
.
They cannot control AI, but they want to control people? …so better to concentrate on the [unpredictable] former, and use modern cyber-security capabilities to keep the former in check.
…or is that an impossibility? …and if so, why so?