How Has AI Changed the Mobile Threat in the Last 12 Months?
In the past twelve months, AI has changed both how attackers reach employees on mobile and how hard their tools are to detect. Zimperium's 2026 Global Mobile Threat Report shows phishing on employee mobile devices up 380% since January 2025 and spyware on nearly one in ten devices. AI use inside mobile apps grew fourteen-fold on Android, often without review.
The report, produced by Zimperium's zLabs research team, argues that AI has brought forward mobile threats that were expected five or more years from now. Here is what has changed, and what it means for organisations and the partners who support them.
Why is mobile one of the least protected parts of the enterprise?
Mobile devices now carry much of an organisation's identity and access, yet few organisations have the visibility of them that they have of laptops and servers. Employee phones have become the main hub for email, collaboration tools, authenticator apps and one-time passcodes. That makes them a high-value target for anyone trying to get into corporate systems.
The average employee phone holds around 80 apps, and only nine of them are for work. The rest are personal and unmanaged, and corporate security tools cannot see them, yet they share the same storage, clipboard and network as the apps used for email and authentication. The report describes each mobile device as four attack surfaces at once: the operating system, the apps, the networks it joins and the person using it.
How much has mobile phishing grown?
Phishing events on employee mobile devices have grown 380% since January 2025. zLabs detected more than 2.5 million phishing attacks on employee devices over the past twelve months. The number of devices on which an employee clicked a malicious link rose by 110% in 2025 compared with the previous year. Citing Verizon's 2026 Data Breach Investigations Report, the report puts the success rate of mobile phishing 40% higher than that of email.
Most of these attacks never pass through an email gateway. SMS accounts for 39% of all mobile threats and 70% of mobile phishing attacks. Impersonation of messaging apps such as WhatsApp and Telegram rose 258%, and Microsoft data cited in the report shows QR code phishing grew 146% quarter on quarter in Q1 2026. Malicious PDFs rose from 19% to 29% of payloads between January and March, a tactic zLabs links to getting past the filtering built into iOS and Android.
Why can't employees spot AI-generated phishing any more?
AI has removed the warning signs that security awareness training teaches people to look for. Research cited in the report estimates that 86% of phishing campaigns now contain AI-generated elements. It also estimates that AI-written lures are around 4.5 times more convincing than those written by people. zLabs found that brand-specific targeting became 5.4 times more precise year on year as attackers concentrated on high-value campaigns.
On a phone, staff have even less to go on. A small screen makes a suspicious link harder to read, and a touch interface offers no easy way to check where a link leads before tapping it. The report concludes that training built around spotting poor grammar and obvious red flags does not carry over to the phishing that arrives today.
How common is spyware and mobile malware on employee devices?
zLabs data shows riskware on more than one in four devices, nearly three times higher than a year ago. Spyware is on nearly one in ten devices, up more than fourfold, and trojans sit on roughly one in twenty. Sideloaded apps installed from outside the official stores appear on 22% of Android devices and 14% of iOS devices.
Commercial spyware has grown into an established market, with more than 500 vendor entities across 40 countries. Google's threat intelligence research found that in 2025, for the first time, more zero-day exploits were attributed to commercial spyware vendors than to nation-state groups.
Banking malware has also moved on. Zimperium tracked 34 active malware families targeting 1,243 financial brands across 90 countries. Many of them now take over transactions on the device itself, and some use AI to rewrite their own code each time they run, which defeats detection based on signatures.
What risks do work and personal apps introduce?
Many apps employees rely on have added AI features since IT approved them, and few organisations go back to review them. AI use inside mobile apps grew fourteen-fold on Android and seven-fold on iOS, with personal app categories such as food delivery and travel among the fastest adopters.
When zLabs assessed top business apps on the App Store and Google Play, it found that 63% of Android apps and 52% of iOS apps make network connections without proper encryption or certificate validation. It also found that 17% of Android apps and 27% of iOS apps call APIs hosted in sanctioned or trade-restricted countries. PII leakage affects 17% of Android apps and 38% of iOS apps. On a phone synced to corporate directories and CRM, that can include the contact details of colleagues, customers and partners.
Where do MDM and existing security controls stop?
MDM shows which apps are installed and which versions they are running, but it cannot show what those apps do, what they access or what they send off the device. App containers isolate corporate data only for as long as the operating system underneath them holds. A rooted or jailbroken phone, or one running spyware with system-level access, breaks that isolation. DLP controls that block copying between managed and personal apps have no effect on an unmanaged personal device.
What are regulators asking organisations to do about mobile?
Governments have started to treat mobile security as a requirement. The UK's anti-fraud strategy names securing mobile devices and apps as a priority, including defences against AI-generated social engineering on mobile. France has published guidance on securing mobile devices. In the US, the Defense Information Systems Agency now mandates mobile threat defence on iOS and Android.
For organisations in scope of NIS2, asset management, access control and cyber hygiene training are all among the expected risk management measures. The phones connecting to corporate systems form part of that picture.
What does this mean for MSPs and resellers?
The report gives partners a straightforward way to open a mobile conversation with customers. The starting questions are what is running on the phones connected to their systems, which apps have changed since they were approved, and whether anyone would know if a device were compromised. Partners already delivering email security, endpoint protection or awareness training can extend those services onto the device itself.
Join us on Tuesday 17th November, 2pm.
The AI-Powered Mobile Threat: What’s Changed and What to Do Next
Andy Grayson at Zimperium, joins Renaissance to unpack the findings of the 2026 Global Mobile Threat Report and what they mean for organisations trying to protect employees, data and access on mobile devices.
The session will cover why mobile phishing is rising so quickly and how AI is making phishing more convincing and malware harder to detect. It will also look at how work apps, personal apps and the AI features built into them are widening the mobile attack surface, and where mobile threat protection sits alongside MDM, email security, endpoint security and awareness training.
The session is relevant both to IT and security leads responsible for employees using mobile devices for work and to MSPs, MSSPs and resellers supporting those organisations.

