Recognized for advancing real-world AI safety through continuous adversarial testing of AI systems
TEL AVIV, April 10, 2026 — Adversa AI announced today that it won the Artificial Intelligence Excellence Awards 2026 in the Safety and Alignment category. The award is judged by leading AI experts and business executives and recognizes companies with real-world impact in AI innovation. The award highlights Adversa AI’s approach to AI safety — focusing not only on theoretical alignment, but on practical, real-world validation of AI behavior under adversarial conditions. Adversa AI was selected for its platform for continuous adversarial testing of AI systems. The platform helps organizations identify risks such as prompt injection, model manipulation, unsafe agent behavior, and unintended actions before deployment. Adversa AI leads the CoSAI Agentic AI Security workstream and serves as a core member of OWASP AIVSS. The platform maps testing coverage to the frameworks its team helps develop. The company’s AI security research has been covered by The Wall Street Journal, Wired, TechCrunch, and Bloomberg. The team created SecureClaw, one of the most widely adopted open-source security frameworks for AI agents.
With quick enterprise adoption of autonomous AI agents, validating AI behavior under real-world conditions has become a board-level concern. Alignment techniques applied at the model level don’t account for how systems behave when inputs are manipulated, tools are misused, or agents chain actions across external systems, creating critical safety gaps that go undetected until exploitation.
“AI safety can’t be validated in isolation from real-world threats. Alignment isn’t just about intent. It’s about how systems behave under pressure, when inputs are manipulated, or when agents interact with tools and external systems. We talk to our peers at CoSAI Agentic AI Security workstream, we talk to our customers, and we see this pattern every day. The standards bodies we contribute to are merging safety and security into a single discipline. This award recognizes that shift. We need to test AI systems the way attackers would. Only then can we understand whether they’re aligned and safe to deploy,”
said Alex Polyakov, Founder and CTO of Adversa AI
The platform lets organizations continuously evaluate AI systems for prompt injection and manipulation risks, unsafe agent actions and decision-making, tool misuse and unintended execution paths, and multi-step workflow vulnerabilities that emerge under adversarial pressure.
Assessments map to OWASP AIVSS, NIST, and CSA standards, letting organizations move from theoretical safety to testable alignment outcomes. The platform brings continuous AI red teaming to enterprise-scale environments in financial services, insurance, and government.
Learn more about the Adversa AI platform.
Full technical guide to Agent Goal Hijack, the #1 risk in the OWASP Agentic Top 10. Explore the attack surface, practical risks, attack examples and practical defense frameworks.
