As generative AI expands the capabilities of hackers and cybercriminals—from deepfaking CEOs to forging receipts—OpenAI is stepping up to support defenses against these advanced threats. The AI giant has made its first foray into cybersecurity, co-leading a $43 million Series A funding round for New York-based startup Adaptive Security, alongside venture capital firm Andreessen Horowitz, the company announced Wednesday.

Founded in 2023, Adaptive Security focuses on simulating AI-driven social engineering attacks to train employees in spotting and resisting them. Its platform mimics phishing attempts via texts, emails, and even spoofed phone calls—like a fake voice message from a company CTO requesting sensitive information. These simulations help organizations identify their most vulnerable areas and reinforce cybersecurity awareness.

“AI has made social engineering attacks easier than ever,” said Adaptive Security CEO and co-founder Brian Long, who previously founded mobile ad startup TapCommerce (acquired by Twitter for over $100 million) and the ad-tech firm Attentive, which reached a $10 billion valuation in 2021.

Adaptive already serves over 100 enterprise customers, and Long credits strong user feedback as a key reason behind OpenAI’s interest. The new funding will primarily be used to hire engineers and further develop the platform in what Long calls the AI “arms race” against increasingly sophisticated cyber threats.

The investment comes as other startups race to tackle the rising tide of AI-enabled cyber risks. Cyberhaven recently raised $100 million to prevent misuse of generative AI tools like ChatGPT, while deepfake detection startup GetReal secured $17.5 million last month.