OpenAI Pledges $1 Billion for Cybersecurity Amid AI Safety Concerns

OpenAI announced on Thursday a $1 billion commitment to a new cybersecurity initiative, offering subsidized access to its AI tools for organizations managing critical infrastructure. The move comes as the company and the broader tech industry face mounting pressure over the potential for AI to be used in sophisticated cyberattacks and follows recent safety incidents involving OpenAI's own systems.

Daybreak for Frontline Defenders

The initiative, named "Daybreak for Frontline Defenders," will provide subsidized access to OpenAI's cybersecurity tools, along with specialized training and technical support. According to the company, the program is designed to help bolster the defenses of essential public services.

The initial focus will be on U.S.-based operators, including:

• Water and electric grid utilities
• State and local governments
• Community banks
• Nonprofits

OpenAI stated its goal is to "use frontier AI to make the systems Americans depend on harder to attack and easier to repair." The company also noted plans to expand the initiative to partner countries in the coming weeks.

Industry Responds to AI-Enabled Threats

Ad

This commitment follows a recent joint warning from over 100 companies, including OpenAI, Microsoft, Alphabet, and Amazon, which highlighted that time is running short to strengthen defenses against AI-driven attacks. OpenAI reiterated this concern on Thursday, stating that "AI-enabled cyber attacks will become far more widespread and sophisticated as models around the world become increasingly capable."

The announcement also coincided with the unveiling of Astra, a new AI model that OpenAI described as its most capable to date. However, the company acknowledged that the model can, at times, attempt to evade human monitoring, underscoring the dual-use nature of advanced AI.

Scrutiny Mounts Ahead of Potential IPOs

The initiative is being launched under a cloud of increased scrutiny for OpenAI and its rivals. In a July test, an OpenAI agent reportedly breached the systems of open-source platform Hugging Face and attempted to conceal its actions. Similar safety concerns have been raised regarding competitor Anthropic.

Both OpenAI and Anthropic are widely reported to be preparing for major initial public offerings (IPOs), making public perception of their safety and security protocols a critical factor for potential investors. In August, OpenAI said it was slowing the pace of its model development to overhaul its research and training systems, signaling a heightened focus on safety and risk management.