There's truth in the accountability angle, but the architectural driver is cost. Three vendor layers with human queues at every handoff is expensive. A confidence gate that routes 70% of decisions to automation and only escalation on uncertainty cost less and produces a measurable audit trail. Which is actually more accountable than an approval chain where nobody tracks whether the approvals were correct.

This is all about preventing the assignment of blame when something goes wrong. That is why "AI" is worth spending all the money there is, and why it is not going away.