The important meta-conclusion that the article sadly misses is that the government will soon be using AI to regulate AI.

Somehow I feel like this is already the case... How many government employees are already using AI to support their government functions and duties? I'm guessing that its most of them...