OpenAI has launched a new program dubbed Daybreak for Frontline Defenders. The initiative aims to provide organizations with the tools they need to secure themselves against increasingly powerful AI-based threats and attackers. But it also goes beyond that, providing companies and public bodies with hands-on support to ensure their services are as hardened as possible in the new era of AI hacking.
The project includes $1 billion in funding that provides “subsidized access to Daybreak cyber models and products, training, technical support, and partnerships in the United States and internationally,” OpenAI revealed. The idea is to offer AI tools that help fortify providers that carry “enormous responsibility without the resources of the world’s largest companies.”
This will at first take the form of a US-based program called Daybreak for America. This is focused on protecting critical infrastructure within the US and includes a new pilot with the Multi-State Information Sharing and Analysis Center (MS-ISAC). The goal is to help both state and local outfits protect the resources that Americans use every day, from water supplies to schools and hospitals.
Organizations and public bodies that team up with OpenAI will be able to “review legacy code, analyze suspicious activity, identify and validate vulnerabilities, prioritize the most serious risks, and develop and test fixes,” OpenAI announced. The initiative will be opened up to other nations in the coming weeks.
Protecting critical infrastructure
The threat posed by AI tools has been a hot topic of discussion in recent months. OpenAI itself recently published an open letter signed by some of the world’s leading tech firms, including Google, Anthropic and Microsoft, calling for a collective response “to raise security standards and find new solutions to emerging cyber threats.”
In announcing Daybreak for Frontline Defenders, OpenAI predicted that “In the coming months, AI-enabled cyber attacks will become far more widespread and sophisticated as models around the world become increasingly capable.” The firm added that “Frontier AI can help defenders move faster,” with its new program being a key part of that strategy. The goal is not to find more vulnerabilities but to resolve them more quickly, OpenAI said.
It follows the launch of OpenAI’s Daybreak program earlier this year. That enabled verified cyber defense firms in the public and private sectors to harness AI tools for defensive purposes. With Daybreak for Frontline Defenders, OpenAI has expanded that to include hands-on tech support and broader access for relevant parties.
Daybreak for Frontline Defenders has emerged at a time when critical infrastructure continues to be targeted by threat actors. Recently, we’ve seen the FBI shut down domains linked to a Chinese state-sponsored hacker and Australian police arrest two men suspected of links to the TeamPCP threat group – in both cases, the attackers had struck bodies linked to critical infrastructure.
It also lands at the same time as the launch of OpenAI’s latest Astra AI model. This has seen a controlled rollout after OpenAI’s own testing revealed that Astra had found and exploited two zero-day vulnerabilities during its development. That leaves OpenAI on both sides of the problem it describes, building the frontier capability and subsidizing the defense against it.
OpenAI ended its announcement of Daybreak for Frontline Defenders with an appeal for relevant bodies to work with it going forward. “The opportunity now is to make sure the advantages frontier AI can give defenders reach beyond the largest companies and best-resourced security teams – and into the communities and institutions whose security affects millions of people,” it declared.