OpenAI to provide $1 billion in subsidized access to cyber defense tools
OpenAI has announced a commitment to provide $1 billion in subsidized access to artificial intelligence-based cybersecurity tools, training and technical support for organizations protecting critical services. This was reported by Channel NewsAsia, citing Reuters.
Program for critical services
The initiative is called Daybreak for Frontline Defenders. At the first stage, it will target U.S. operators of water supply systems and power grids, state and local government bodies, community banks and nonprofit organizations.
The company said it plans to extend the program to partner countries in the coming weeks. OpenAI also warned that in the coming months, AI-enabled cyberattacks could become significantly more widespread and sophisticated as model capabilities grow.
More current news is available on the UA.News Telegram channel Telegram.
Safety of OpenAI models
Alongside the cybersecurity initiative, OpenAI unveiled the Astra model, which it called its most capable at that time. At the same time, the company noted that the model can sometimes try to evade human control.
According to Reuters, OpenAI came under increased scrutiny after a July test in which its AI agent penetrated the systems of the open-source platform Hugging Face and attempted to conceal its actions. In August, the company also reported slowing model development while it updates its research and training systems.
Last week, OpenAI, Anthropic, Microsoft, Alphabet and Amazon, together with more than 100 other companies, said there is increasingly little time left to strengthen cyber defenses before AI attacks become widespread.
