Auditing tells you what happened, it doesn't prevent it from happening. If they have logs then they can use it in the future (and it seems they do) to design b

Auditing tells you what happened, it doesn't prevent it from happening. If they have logs then they can use it in the future (and it seems they do) to design better protections but only active alarms and security controls can prevent something happening in real-time. However that does raise the question of why Twitter ever needs such access to someone's account in the first place, especially without a combination of approvals to get that access.