As reported by The Hacker News, SOCRadar's Threat Research Unit has published findings on AnonyMousKIT, a credit-metered phishing-as-a-service platform that orchestrates multi-channel lures — including AI-generated voice calls — against owners of stolen Apple devices. The goal: extract device passcodes, Apple ID credentials, and live 2FA codes to defeat Activation Lock. The platform's recovered Vapi account alone contained 200 call records and 55 transcripts across five personas, all impersonating "Alice from Apple Support."

Threat Alert: As reported by The Hacker News, SOCRadar's Threat Research Unit has published findings on AnonyMousKIT, a credit-metered phishing-as-a-service platform that orchestrates multi-channel lures — including AI-generated voice calls — against owners of stolen Apple devices.

Why This Matters Beyond Apple Ecosystems

The AnonyMousKIT disclosure is significant not because Activation Lock bypass is novel — theft-related social engineering has plagued Apple users for years — but because of the operational maturity and cost economics on display. At approximately 9.6 cents per AI voice call, the barrier to launching convincing vishing campaigns has collapsed. A threat actor can now dial 1,000 victims for under $100, with automated transcripts, multilingual personas, and CRM-style outcome tracking. This is not a phishing kit; it is a vertically integrated criminal SaaS.

The recovered lures cited each device's internal Apple model identifier and live Find My status — data pulled directly from the stolen handset. That level of personalization makes standard "verify the caller" advice insufficient when the caller already knows details only Apple should possess.

The AI Voice Platform Abuse Problem

The use of Vapi — a legitimate commercial voice AI platform — illustrates a broader supply chain risk that few organizations are addressing. Threat actors are renting enterprise-grade conversational AI infrastructure the same way they rent cloud compute or CAPTCHA-solving services. The personas recovered were consistent, scripted, and capable of reading back entered digits for confirmation — a UX detail that increases victim compliance. This is not a robocall; it is an interactive negotiation.

For defenders, this means traditional vishing indicators — awkward pauses, background noise, non-native accents, scripted rigidity — are increasingly unreliable. AI voice agents can maintain natural cadence, adapt to responses, and operate around the clock across time zones without fatigue.

Who Is at Risk

The AI Voice Platform Abuse Problem
Stolen device owners — particularly in regions with high device theft rates. Brazil accounted for 179 of 200 recovered calls, but the multilingual persona set (English, Spanish, Portuguese) signals broader geographic ambition.
Enterprise mobility programs — organizations issuing Apple devices to employees should assume that a stolen corporate handset will be targeted through the employee, not just through technical bypass attempts.
AI service providers — Vapi and similar platforms face reputational and potentially regulatory exposure if their services are systematically abused for fraud at scale.

Shield53 Recommendations

  • Update security awareness training immediately. Move beyond "Apple will never call you" messaging. Teach users that attackers now possess device-specific details (model identifiers, Find My status) and can simulate interactive support conversations. Emphasize that no legitimate support process — from any vendor — requires reading back a device passcode over the phone.
  • Issue standing guidance to employees with corporate Apple devices. If a device is lost or stolen, employees should contact IT through pre-established internal channels only, not respond to inbound communications referencing the device. Document this in incident response runbooks.
  • Enable Apple's Stolen Device Protection on all managed and BYOD Apple devices. This introduces a security delay for sensitive actions when the device is away from familiar locations, buying time against social engineering-driven credential theft.
  • Monitor for AI voice platform abuse in your threat intelligence program. Track mentions of commercial voice AI services in phishing infrastructure. If your organization runs a customer support operation, assess whether your own voice AI providers have anti-abuse controls.
  • For MDM administrators: Enforce Activation Lock management through Apple Business Manager / Apple School Manager so that Activation Lock is tied to your institution's Apple ID, not the employee's personal account. This reduces the value of social engineering the individual user.

The AnonyMousKIT case is a preview of where consumer-facing fraud is heading. The intersection of stolen device data, legitimate AI voice infrastructure, and SaaS-style criminal operations creates a threat model that awareness training alone cannot solve. Defenders need technical controls, identity-bound recovery processes, and supplier due diligence on the AI platforms their own organizations and ecosystems depend on.