The Pentagon's Secretive AI Hacking Venture: A Deep Dive
The U.S. military is quietly investing millions in a cutting-edge AI startup, Twenty, to develop offensive cyber capabilities that could reshape the landscape of cyberwarfare. This move raises important questions about the ethical and strategic implications of AI in warfare.
The AI Hacking Startup: Twenty
Twenty, a stealth startup based in Arlington, Virginia, has secured a contract worth up to $12.6 million with the U.S. Cyber Command. This contract is significant because it marks a rare instance of a VC-backed AI offensive cyber company landing such a contract. The company's investors include In-Q-Tel, Caffeinated Capital, and General Catalyst, indicating a high level of confidence in their capabilities.
Twenty's website boldly states its mission: "Transforming workflows that once took weeks of manual effort into automated, continuous operations across hundreds of targets simultaneously." They claim to be "fundamentally reshaping how the U.S. and its allies engage in cyber conflict."
Job postings reveal a focus on developing advanced offensive cyber capabilities, including attack path frameworks and AI-powered automation tools. They plan to utilize open-source tools like CrewAI, which enables the management of multiple autonomous AI agents working together.
A Team of Veterans
Twenty's executive team is composed of former military and intelligence professionals. CEO and co-founder Joe Lin, a former U.S. Navy Reserve officer, brings expertise in product management from his time at Palo Alto Networks. CTO Leo Olson and VP of Engineering Skyler Onken also have extensive experience in national security and cyber operations.
The Growing Role of AI in Cyber Warfare
The U.S. isn't alone in its pursuit of AI-powered hacking capabilities. AI giant Anthropic recently revealed that Chinese hackers were using its tools for cyberattacks, showcasing the potential for AI to significantly enhance hacking efforts.
The U.S. Defense Department has awarded contracts to OpenAI, Anthropic, and xAI for unspecified "frontier AI" projects, further fueling speculation about the potential involvement of these companies in offensive cyber operations.
Comparing Automation Levels
While Twenty aims for simultaneous attacks on hundreds of targets, another contractor, Two Six Technologies, has received substantial AI offensive cyber contracts but focuses on assisting humans rather than replacing them. Their AI, developed under the IKE project, can initiate attacks if success probabilities are high.
The Dual Nature of AI in Cybersecurity
AI is predominantly used defensively, with startups like Tenzai leveraging it to identify vulnerabilities in software. However, the dual nature of AI in cybersecurity raises concerns about the potential for misuse and the need for ethical guidelines and regulations to govern its development and deployment.
