KYC automation vs. the moving target of new fraud vectors

Been thinking a lot about the push towards fully automated KYC processes, especially for faster onboarding in fintech. While the efficiency gains are undeniable, it feels like we're in a perpetual arms race. Every time a new AI-driven solution promises to cut verification times down to seconds, a new fraud vector emerges that bypasses existing checks. It's not just about identity documents anymore; synthetic identities, deepfakes, and more sophisticated mule networks are making it incredibly challenging. How are others balancing the need for speed and customer experience with the ever-evolving landscape of AML risks? Are the current regtech solutions keeping pace, or are we always a step behind?