Horizon3 launches Adversarial Intelligence Frontier Lab

0

Horizon3 has announced the launch of its Adversarial Intelligence Frontier Lab, a research initiative focused on developing “closed-loop” systems that use evidence from production-safe NodeZero attack graphs to identify and close real-world attack paths at machine speed.

In the release, the company positioned the lab’s approach as distinct from AI security efforts that rely primarily on sensor telemetry or testing against digital twins, saying its work will begin from “verified, end-to-end attack paths” that have already succeeded against live systems. It said research will focus on keeping attack-and-defence systems “safe, observable, and controllable” in production environments.

“We’ve spent years collecting the attack-and-defence data required to research these systems. The next step is to use that verified attack-path evidence to improve how offensive and defensive agents work together in environments that actually matter,” said Snehal Antani, CEO and co-founder of Horizon3.

Research leadership

Dr. Z. Berkay Celik will serve as the inaugural director of the Adversarial Intelligence Frontier Lab. Horizon3 said Celik is an Associate Professor of Computer Science at Purdue University and co-director of the Purdue Security Laboratory, with research focused on vulnerability discovery, attack analysis, and the design and evaluation of defences for agentic and autonomous systems.

The company said the lab builds on Horizon3’s published research into AI attacker behaviour, cyber deception, and autonomous defence, citing three papers hosted on arXiv: “Honeyquest for LLMs: Rethinking Cyber Deception for AI Attackers”, “Closing the Sim-to-Real Gap: An Evaluation Framework for Autonomous Cyber Defense Configuration of Commercial EDR”, and “Stable Agentic Control: Tool-Mediated LLM Architecture for Autonomous Cyber Defense”.

Four research lines

Horizon3 said the lab will initially pursue four lines of research: AI-enabled attacks and adversarial behaviour; adaptive defence and attack-and-defence learning loops; safe, controllable, and measurable autonomy; and cyber resilience and operational readiness, including training and workforce preparation.

The company also said the lab will accelerate training of “customer-specific cyber world models” intended to support shared context and long-term memory for red and blue agents across attack and defence cycles.

Claims on training data and scale

Horizon3 said its customer base of more than 7,000 organisations generates “diverse, labelled training data” from “hundreds of thousands” of production-safe NodeZero tests. It estimated it is collecting 30 billion tokens of anonymised, aggregated, labelled attack-defend training data per day, including attacker commands executed, whether commands succeeded or failed, defensive tool responses, and configuration context behind the firewall.

Critical infrastructure focus

Horizon3 said the lab will oversee its Cyber Resilience Centres of Excellence focused on critical infrastructure, aiming to translate research into improvements for essential systems where downtime or compromise may carry national or economic consequences.

Share.