Attackers are increasingly using AI-Apple phishing to trick victims not only through email or SMS, but also via phone calls with an artificial voice. In a new investigation, researchers describe a platform built specifically for “unlocking” stolen Apple devices—not by cracking real security, but by getting people to hand over the required information step by step.
At the core is Activation Lock: a feature that links devices to an Apple ID and makes the device unusable after that theft until the owner’s account is removed. Researchers show how fraudsters bypass Activation Lock through interception via social engineering, with the end goal of obtaining a live 2FA code.
What exactly is AI-Apple phishing?
According to researchers, this isn’t a single standalone phishing campaign, but a phishing-as-a-service (PhaaS) platform. With a credit system, criminals can deploy channels (such as email, SMS, WhatsApp, and voice calls) using a single set of “victim data.” The campaign is designed to first convince a victim, and then immediately request the right sensitive information.
The latest twist is the AI-voice channel. Researchers describe how voice agents are configured to call with a persona that resembles Apple Support. This creates a route where victims aren’t only sent to a webpage, but are also guided by phone.
Activation Lock and why victims are targeted
Activation Lock was introduced with iOS 7 and locks a stolen handset to a specific Apple ID. In practice, that means the theft yields little without removing that link.
That’s why scammers focus on the owner. In the lures, the device is approached as if it has been “found” or “flagged,” with the aim of getting the user to provide the data needed for account access.
How scammers make victims hand over information step by step
The messages and calls request information in a sequence that matches the Apple support flow—only twisted for fraud. Victims are asked to:
- provide the 4- or 6-digit device passcode,
- then Apple ID details,
- and finally a live 2FA code received at that moment.
Important detail: Apple emphasizes that the company never asks for a password, device passcode, or 2FA code to provide support. If, in a conversation or via a page, you’re required to share that information, it’s almost always a red flag.
Why AI-voice calls are so effective
One of the most documented methods in the investigation is the AI-voice channel. Researchers mention a set of call recordings and transcripts recovered from the speech platform used by the attackers.
The call is designed to make the victim actively confirm that they are the “owner.” Then comes the moment when the agent reads out and checks the requested digits—for example by asking the passcode to be repeated. Next, the victim is guided through a scenario explaining that someone would have gone to an Apple Store to remove Activation Lock, after which they’re asked whether a “recovery” link was received.
This kind of interaction makes it harder to recognize the scam afterward as “classic phishing,” because the victim feels like they’re part of a real-time process.
Channels, pricing, and scale: from one victim record to multiple routes
The platform researchers follow (as they label it in their report) appears to be “credit-metered”: using different channels costs credits. The investigation lists costs and channels, including email, SMS (priced per sender ID), WhatsApp, recorded voice calls, and an AI-voice agent.
It also describes how a single “victim record” can run across multiple channels. That means: using the same core information, a victim can be approached in multiple ways at the same time or sequentially, increasing the chance of success.
What does a victim page look like?
The lures send victims to web pages that look like Apple-branded content. The investigation reports capture pages with a “token” in the URL. Those pages display an animated map that suggests the location of the device that has supposedly been “found.”
For the fraud, the most important part is that the pages are designed to push users to the next step: the passcode, Apple ID login, and ultimately the 2FA code.
What works for defense? Not just technology, but behavior too
Researchers highlight a specific remediation you can implement both as an organization and as a user: move high-value Apple IDs, where possible, to physical hardware security keys. This drastically limits the attackers’ ability to intercept real-time 2FA.
In addition, awareness remains crucial. If, during support—especially by phone—you’re asked for sensitive access (such as passcodes or 2FA codes), the odds are high that you’re dealing with AI-Apple phishing.
Extra context: the same ecosystems and infrastructure
These campaigns don’t exist in isolation. The investigation references earlier analyses of vishing services using text-to-speech, as well as domain observations suggesting a broader “unlock-kit” ecosystem. In that wider context, it’s emphasized that the combination of technical tooling and social engineering makes the difference: attackers can make money at scale through phone-based theft.
It also stands out that researchers identify multiple storefronts that were launched at the same time and show similarities in accounts. This suggests a single organized actor or a tightly connected network, although that doesn’t automatically mean the same operator behind every domain.
Practical checklist: how to recognize this kind of fraud
Use these points as a quick check when you receive “Apple Support”-like contact about a stolen or found device:
- Are they asking you for a passcode or 2FA code? Stop. Apple doesn’t do that.
- Do you get a live conversation with a “Support agent”? Treat it as suspicious until proven otherwise.
- Are you being sent to a token URL or a “recovery” page? Never enter your login or codes via links from a message.
- Are location claims made (like “a map with Find My”)? A plausible UI isn’t proof.
For additional background on how automated social engineering and vishing work in practice, you can also read how AI-driven vector dynamics puts security under pressure—for example in analyses of vulnerabilities and deception in other chains. A relevant, broader perspective is available in our article about Application security in the AI-tijdperk.
What you can do now if you’re a victim (or think you are)
Say you may have shared a passcode or 2FA code, or you visited a page that requested a login. Take action immediately:
- Stop sharing information further. Don’t send any additional codes.
- Check your accounts for abuse. Change your login details where needed via the official Apple route.
- Limit impact with a strong second factor. If your device or account settings allow it, switch to hardware security keys.
Also contact Apple using official channels. Don’t search for information based on the link or instructions from the scam—use the path you already know.
If you want to understand how attackers structure their campaigns and why that matters for defense, it helps to look in parallel at earlier attacks where social engineering and infrastructure come together, such as Iran-linked hackers and impact on infra. That kind of context shows why it’s essential to address both the human and technical sides.
Conclusion: AI-Apple phishing makes theft more profitable
The case described shows how fraudsters use Activation Lock as the starting point for an attack, but ultimately “open the lock” using another method: persuading people to hand over the right codes. With AI-Apple phishing—especially via AI-voice calls—social engineering becomes more interactive and therefore more convincing.
The best defense is a combination of vigilance (don’t share passcodes or 2FA codes), fast action when in doubt, and strengthening account security with hardware security keys where possible. That reduces the chance that an AI-driven scam can bypass your access.
Source: https://thehackernews.com/2026/08/fake-apple-support-ai-calls-target.html
