Today we discuss the evolution of AI tools in cybercrime, highlighting how these malicious platforms have advanced from early, unsophisticated chatbots to complex, multi-modal systems. We describe the capabilities and distribution of key dark web AI tools like WormGPT, FraudGPT, EvilGPT, and the cutting-edge Xanthorox AI, detailing their uses in generating phishing content, developing malware, and enabling fraud. We also assess the threat level posed by these "AI-as-a-service" offerings, noting their ability to lower the technical barrier for cybercriminals and suggesting countermeasures for defense. Ultimately, we emphasize that AI has become a double-edged sword, necessitating proactive cybersecurity strategies that incorporate AI-driven defenses to combat the rising tide of AI-augmented attacks.