iProov report warns of soaring iOS injection attacks

iProov has published its 2026 Threat Intelligence Report, which says injection attacks targeting iOS devices rose 1,151% in the second half of 2025.

The report examines how generative AI is reshaping identity-based attacks, focusing on digital identity verification, impersonation and fraud methods used against organisations. It draws on data from iProov's Security Operations Centre, along with external threat intelligence, dark web monitoring, red-team testing and biometric security research.

The sharp rise in iOS-targeted injection attacks was one of the clearest shifts identified in the study. Attacks on Apple's mobile platform rose just 14% in the first half of 2025 before accelerating sharply later in the year, contributing to a 741% annual increase.

Injection attacks involve feeding false or manipulated material directly into systems instead of presenting a live person to a camera or sensor. The report suggests methods once seen as limited or experimental are now being repeated at much greater scale.

Enterprise Risk

Another theme is the spread of deepfake impersonation into routine workplace activity. According to iProov, synthetic identities are moving beyond formal identity checks and into day-to-day corporate processes, especially video-based interactions.

That finding reflects broader concern across the cyber security sector about the use of AI-generated audio and video in fraud, social engineering and executive impersonation. The report cites research from the Ponemon Institute showing that 41% of organisations have experienced deepfake attacks targeting executives, while a Gartner study found that 37% of cyber security leaders had encountered deepfake incidents during video calls.

Advances in image-to-video generation have lowered the barrier to producing convincing synthetic personas from limited source material, the report says. It points to widely available AI tools that can quickly generate realistic visual content, making impersonation easier across internal business workflows.

The report also links identity-related attacks to broader access and operational risks. It says recent incidents affecting companies including Marks & Spencer and Jaguar Land Rover show how weaknesses in identity and access controls can leave organisations exposed to disruption after a successful impersonation or social engineering attempt.

Regional Patterns

Geography is another major focus. The report says identity fraud is becoming more global, with Southeast Asia emerging as a testing ground for new attack methods before they spread elsewhere.

Southeast Asia recorded a 720% increase in attacks in the third quarter of 2025. According to the report, criminal groups in the region have used techniques including virtual camera attacks and stolen know-your-customer identity packages, with those methods later adopted in other regions, particularly Latin America.

This suggests a more networked and transferable fraud ecosystem, where successful tactics can move quickly across borders and sectors. Financial institutions and digital platforms are among the likely targets because they depend heavily on remote onboarding, account access controls and transaction approval systems.

Dr. Andrew Newell, Chief Scientific Officer at iProov, said: "Identity is becoming the new battleground in cybersecurity. Generative AI is allowing attackers to industrialize digital impersonation at scale. To defend against this, organizations must be able to establish genuine human presence in digital interactions to ensure trust and security."

Detection Shift

The report argues that organisations can no longer rely on fixed approaches to identity verification and authentication. Older methods, it says, assume a stable threat environment and depend too heavily on static testing, leaving security teams less prepared for rapidly changing AI-driven attack techniques.

Instead, companies need continuous identity threat detection that can adapt as attack patterns evolve, according to iProov. The report also points to standards including NIST SP 800-63-4, CEN/TS 18099 and FIDO Face Verification Certification as part of a broader response.

The findings add to growing evidence that identity systems are becoming a more contested area of cyber security as generative AI tools improve and spread. The report's central claim is that attacks are becoming not only more realistic, but also easier to reproduce, distribute and adapt across devices, geographies and routine business interactions.