The article doesn't specify the roles of these four individuals. Usually, the field executors are the first to be arrested; this doesn't mean the operation is over if the masterminds of the syndicate haven't been caught. $140 million is enough to turn a city's population into criminals, not just 67 accomplices.
Exactly, let's see how big this syndicate is! But whenever they find more people, they don't announce it publicly for a long time. Operation's success could be the reason behind it. They don't want to share something that could affect their operations. Anyway, if we analyze other hacks and criminal syndicates running for years, we will realize that sometimes even 4 people could pull stuff like this if they have access to AI models at an advanced level.
They can automate their tasks easily. And it has been proved by the number of hacks and exploits that since the inception of AI adoption, they are increasing, and this year with new models of Anthropic, things are getting easier.
Everyone is finding vulnerabilities in systems, although this might not be relevant to cyber fraud because these models are also centralized and authorities can backtrack the hackers. But if they are not using AI, then why have their attacks increased dramatically?