When AI Becomes the Criminal: Machine Learning in Modern Fraud
In early June 2025, India's Central Bureau of Investigation announced the dismantling of two illegal call centers and the arrest of six individuals as part of a coordinated action known as Operation Chakra V. The operation, carried out in partnership with Microsoft and Japanese cybercrime authorities, targeted a scam network accused of impersonating Microsoft tech support to defraud citizens in Japan. Coordinated raids took place on May 28, 2025, across 19 locations in Delhi, Haryana, and Uttar Pradesh.
A Global Scam Ring Falls: Inside Operation Chakra V
According to the CBI and Microsoft, the operation targeted a call center network that posed as legitimate technical support representatives to trick victims into granting remote access to their devices or paying for fake services. Japanese citizens, including elderly individuals, were reportedly the primary targets. The scale and cross-border nature of the operation—spanning Indian law enforcement, a major US technology company, and Japan's Cybercrime Control Center—illustrates the increasingly transnational character of modern scam operations.
Where AI Enters the Picture
Coverage of the takedown points to the reported use of generative AI tools in scaling the scam operation, including AI-assisted translation and localization that helped non-native speakers convincingly target Japanese victims. Many observers note that this fits a broader pattern: generative AI is increasingly cited by industry analysts as a tool that lowers the technical and linguistic barriers required to run fraud-as-a-service operations. Supporting industry analysis, including research from TRM Labs, points to examples such as automated phishing content generation, multilingual scam scripting, and deepfake voice cloning as capabilities that could make scam operations more convincing and easier to scale. It's worth noting that many of these AI-specific mechanics are described in adjacent industry commentary rather than confirmed in official CBI or Microsoft statements about this particular case, so readers should treat the AI angle as a contextual and increasingly common feature of scam ecosystems rather than a fully documented element of this specific investigation.
The Scale of the Takedown
Microsoft's Digital Crimes Unit, working alongside Japan's Cybercrime Control Center, reports having taken down approximately 66,000 malicious domains and URLs since May 2024. This figure, drawn from Microsoft's own corporate statements, points to an ongoing and large-scale enforcement effort rather than a single isolated case. A separate but related joint operation between Indian authorities and the FBI reportedly dismantled a call center scam network estimated at $48 million, suggesting a broader pattern of transnational, tech-support-themed fraud operations rather than a one-off incident.
Who's Verifying These Claims?
It is important to distinguish between confirmed law enforcement action and claims that originate primarily from corporate or law enforcement communications. The core facts of Operation Chakra V—the arrests, the raid locations, and the number of dismantled call centers—are attributed directly to CBI statements and corroborated by multiple outlets. The takedown figures regarding malicious domains, however, come primarily from Microsoft's Digital Crimes Unit blog posts and have not been independently audited in public reporting reviewed here. Readers should also note that early coverage of this story appeared alongside loosely related items, including a Coinbase-linked insider data breach involving third-party support contractors and a separate INTERPOL operation resulting in arrests connected to child sexual abuse material. Neither of these is directly tied to the AI-fraud thesis at the center of this article, and they are mentioned here only to clarify what is—and is not—part of the core case.
What This Means for the Next Wave of Fraud
A recurring consumer concern raised by analysts is that generative AI tools continue to reduce the cost and skill required to run convincing scam operations at scale, even as law enforcement agencies develop new models of cross-border cooperation. The CBI-Microsoft-Japan collaboration, along with the separate India-FBI operation, may offer a template for how enforcement agencies and private companies can coordinate against AI-enabled fraud networks. The FBI's consumer guidance continues to emphasize basic recognition tactics for tech support scams, such as unsolicited contact claiming to be from a known technology company and requests for remote device access or unusual payment methods. Still, an open question remains: as generative AI tools become more capable and accessible, some observers wonder whether enforcement efforts—however well-coordinated—can keep pace with the speed at which AI-assisted scam ecosystems can adapt and scale.