Martin Zugec

What’s coming next in cybersecurity? Hear from industry experts as they break down emerging attack trends, the impact of AI on both attackers and defenders, evolving email and endpoint threats, and what security operations teams are seeing firsthand.

You’ll leave with a clearer understanding of where the industry is headed, how to strengthen your defenses, and how to better prepare your customers for what’s next.

And because three hours of cyber talk deserves some perks: come hungry, enjoy the open bar, and stick around for giveaways throughout the session.

This session is based on three real incidents that Bitdefender investigated - actual intrusions, not malware samples reverse-engineered in isolation. Different things get called “AI attacks,” and these real incidents show they are not the same threat. Together they support a classification of external AI-enabled threats, separated by what the model actually does in an operation: supports the operator, produces the artifact, or runs the attack.

SilkParasite, a China-nexus actor, illustrates AI-assisted operations: models supporting reconnaissance, social engineering and tooling. APT36, a Pakistani actor, illustrates AI-generated vibeware - tooling that was model-written, and that proved derivative, bloated and readily detected in the field. The third case, a 3AM ransomware intrusion, illustrates AI-operated activity, where the operation itself was agent-driven.

For each tier we show how to recognize it, what it leaves behind in your telemetry and detections, and how worried you should actually be. Drawing on Bitdefender Labs and MDR data, I’ll give a verdict on each: a gimmick, an experiment, or something you should expect more threat actors to adopt in 2027.