Cybersecurity is at an inflection point, driven by a profound shift powered by artificial intelligence (AI). My recent conversation with Matan Derman, CEO of Apex, reinforced my conviction that AI’s integration into organizations is inevitable\u2014and that new, AI-specific threats require a fundamentally rethought security strategy.<\/p>\n\n\n\n
Derman, whose distinguished background includes prestigious service in Israeli cybersecurity units and Stanford Business School, brings deep insights into why organizations must urgently rethink security in the AI era. Together with his co-founder, Tomer Even, a partner with an equally impressive track record, they recognized early the fundamental shifts AI was bringing about. “It was clear that the world was changing\u2014AI tools had existed for years, but only recently did they become universally accessible,” Derman explained.<\/p>\n\n\n\n
The accessibility and ubiquity of tools like ChatGPT and GitHub Copilot have revolutionized productivity but simultaneously opened the floodgates to new vulnerabilities. When Derman discovered organizations actively banning these tools due to security concerns, it became a clear “aha” moment: “Organizations were banning AI tools out of fear\u2014legal, compliance, privacy, and data leakage concerns were overwhelming.”<\/p>\n\n\n\n
This realization resonates deeply with my own experience at Clarity, where I\u2019ve witnessed firsthand the escalating sophistication and frequency of AI-generated cyberattacks. According to a Team8 report, AI-driven phishing attacks alone have increased by more than 2,000% in recent years. Organizations’ traditional security methods\u2014centered around static perimeters and reactive measures\u2014are ill-equipped to manage these dynamic threats.<\/p>\n\n\n\n
One of Derman’s key insights was the recognition of AI as a double-edged sword\u2014introducing immense potential alongside significant threats. Apex\u2019s approach, therefore, revolves around building AI-powered solutions designed explicitly to secure AI-driven workflows. As Derman emphasized, “We evolved from being a security company focused on AI, to becoming a security company for AI that itself leverages AI extensively.”<\/p>\n\n\n\n
During our discussion, Derman highlighted that traditional cybersecurity methods fail to address AI-specific threats adequately, such as “prompt injection” and “jailbreak” attacks. These attacks trick AI models into bypassing their safety measures, revealing sensitive or protected information. “We’ve demonstrated jailbreak techniques even using Morse code,” Derman noted, underscoring the limitless variations attackers now utilize. These tactics demand entirely new approaches to threat detection, making AI-driven defenses essential rather than optional.<\/p>\n\n\n\n