The AI Arms Race in Mobile App Security

Artificial intelligence is changing mobile security by making existing attack and defense techniques faster, more scalable, and more accessible. In this webinar, experts from Keysight and Promon discuss how AI is affecting mobile applications, payments, identity, authentication, application protection, and security evaluation, and what organizations should consider as these capabilities continue to evolve.

The session begins with the attacker perspective. AI can accelerate vulnerability discovery, code analysis, reverse engineering, phishing, impersonation, fraud, and reconnaissance. Tasks that once required significant time and specialist knowledge can increasingly be supported by large language models and purpose-built AI systems. This can lower the barrier for less experienced attackers while helping sophisticated attackers operate more efficiently.

For mobile applications, AI-assisted reverse engineering is an important area of concern. Models can help identify functionality, analyze code, recognize protection mechanisms, and support de-obfuscation. The panel also explores how deepfakes, synthetic voices, and AI-generated content are making impersonation attacks more convincing, placing greater pressure on identity and authentication systems.

On the defensive side, AI is already helping security teams analyze malware, investigate vulnerabilities, review code, reproduce attacks, and automate parts of security testing. However, the speakers emphasize that AI should augment rather than replace expert judgment. Outputs still need to be validated, contextualized, and assessed by security professionals.

The discussion also examines how authentication and application protection may need to evolve. Layered controls such as liveness detection, hardware-backed integrity, runtime protection, contextual signals, and continuous verification can make attacks more difficult and expensive. Runtime protections remain particularly important because preventing attackers from freely executing and manipulating an application can force them toward more time-consuming analysis.

Security evaluation and certification must also adapt. Existing methodologies remain relevant, but the assumptions behind attacker capabilities may need to change as AI reduces the time and expertise required for certain attacks. Evaluators may increasingly need AI-assisted tools themselves to represent realistic attacker capabilities, while maintaining strict controls over confidential source code, product information, and other sensitive data.

The webinar also considers the security of AI-enabled systems themselves, including adversarial inputs, prompt injection, deepfakes, side-channel analysis, and fault injection against AI workloads.

Ultimately, the discussion reinforces that AI does not replace established security principles. Security by design, defense in depth, isolation, strong authentication, runtime protection, vulnerability monitoring, and continuous testing remain critical. AI changes the economics and pace of attack and defense, making it increasingly important for organizations to continuously reassess their security assumptions, testing methods, and protection strategies.