OpenAI Fuels Fight Against Deepfakes with Adaptive Security Investment

Summary

• Investment Announcement: OpenAI’s Startup Fund co-led a $43 million Series A in Adaptive Security, OpenAI’s first cybersecurity investment.
• Technology Development: Adaptive simulates deepfake calls and AI phishing to train staff and score risk.
• Market Impact: The deal reflects rising worry about AI-powered social engineering.
• Strategic Importance: It signals OpenAI’s interest in defenses against misuse of AI.

OpenAI’s First Cybersecurity Bet

OpenAI has made its first investment in a cybersecurity company. On April 2, 2025, Adaptive Security announced a $43 million Series A round co-led by Andreessen Horowitz and the OpenAI Startup Fund. (Earlier versions of this article put the figure at $50 million and described it as an investment by OpenAI alone; that has been corrected.) The startup focuses on attacks that use AI-generated voices, video and text to fool employees.

How Deepfake Scams Work

Deepfakes use AI models to create realistic audio or video that is hard to tell from the real thing. Criminals now use them to impersonate executives on calls, or to make phishing messages far more convincing. The same tools feed misinformation campaigns and personal attacks.

Adaptive Security’s Groundbreaking Approach

Adaptive Security runs realistic simulations of deepfake calls and AI-written email and text phishing, so companies can see how their staff respond. It pairs that with tailored training, triage of reported threats and a risk score for the organization. CEO and co-founder Brian Long said the money would help scale defenses against AI-powered social engineering. (A quote attributed to a “Jane Elwood” as CEO in an earlier version has been removed; she is not the company’s CEO.)

Market Influence and Broader Implications

Deepfakes are a growing risk for media and elections, and for any business that approves payments over the phone. Backing from OpenAI and a16z gives Adaptive money and visibility at a time when these attacks are spreading. (A quote attributed to a named analyst in an earlier version has been removed because we could not confirm it.)

A Collaborative Effort in Cyber Defense

The deal ties an AI developer to a company that defends against AI misuse. That pairing is likely to draw more interest from investors.

Looking Ahead: Challenges and Opportunities

The work is far from done. Deepfake tools improve quickly, so defenses have to keep adapting. Training people to pause and verify before acting remains one of the best protections.

What the Deal Signals

OpenAI’s stake in Adaptive Security shows that AI makers see social engineering as a serious threat. For companies, the practical lesson is simple: assume a convincing voice or video can be faked, and build checks around payments and access that do not depend on recognizing someone.

Frank Jones, CISSP

Frank Jones has loved computers from the age of 13. Frank got his hacking career started when he downloaded a war dialing program that he used to detect dial up modems in his hometown of Chicago. Frank Jones now works as a JAVA coder and cyber security researcher.