Address open challenges in dynamic malware detection

Develop effective approaches for detecting fileless malware residing entirely in memory, AI-powered malware that adapts its behavior to evade detection, and data-poisoning attacks that gradually degrade malware-detection accuracy, while addressing the privacy and ethical challenges of executing untrusted code.

Background

The paper concludes that static API-import analysis cannot observe several runtime or adaptive threats. It specifically identifies fileless malware, behaviorally adaptive AI-powered malware, and attacks that poison training data over time as unresolved challenges for dynamic malware detection.

The authors also note that executing untrusted code to support dynamic analysis creates privacy and ethical complications. Hybrid static–dynamic systems are proposed as a possible direction, but their effective integration is expressly described as technically difficult.

References

Finally, future research must address the open challenges associated with dynamic malware detection, including fileless malware that resides entirely in memory, AI-powered malware capable of adapting its behavior to evade detection, and adversarial threats such as data poisoning attacks that gradually degrade model accuracy.

— Delphi Scanner: efficient and interpretable static malware detection via API sequence modeling  (2609.19900 - Brahimi et al., 17 Sep 2026) in Section 6.5, “Future Work”