"AI-Driven Hacking Is Spreading, and There Is No Time to Respond"...116 Big Tech Firms Call for Global Joint Action [Global AI Briefing]

[Financial News] "Over the next few months, the performance of AI models around the world will continue to improve, and AI-based cyberattacks will become far more widespread and sophisticated. The time available to strengthen cyber defenses is limited."
A total of 116 global Big Tech companies, including OpenAI, Anthropic, Google Experience Center, Microsoft (MS), Amazon Web Services (AWS), and Oracle Corporation, said AI-powered cyberattacks are becoming broader and more powerful. They called for a global coordinated response, including providing defenders with stronger AI tools.
On the 27th local time, OpenAI said in a joint letter titled "A Call for Collective Action on Cyber Defense" that AI cyberattacks will spread beyond the internet to companies and public services that society depends on, including hospitals and water treatment plants.

OpenAI also disclosed recent cases in which AI models bypassed safeguards and penetrated external systems during cyberattack scenarios. In a July internal cybersecurity evaluation of its latest AI model, the company said a research model bypassed restrictions on internet access and compromised parts of OpenAI's internal research infrastructure and the Hugging Face system.
The Big Tech firms also said, "Advances in AI technology can become a weapon for defenders to address vulnerabilities that have accumulated over years," and urged the use of AI for preparedness.
The joint letter urged frontier AI companies to provide powerful AI models and tools to organizations responsible for cyber defense, while governments should supply the resources needed for critical infrastructure such as hospitals and public services.
It also stressed that companies and governments should share cyber threat intelligence and work together to automate vulnerability detection, patching, and security testing with AI.
The goal is not to block AI technology out of fear of cyberattacks, but to actively use it to defend against them.
Unlike earlier AI safety debates, which focused mainly on preventing AI from generating dangerous content, the Big Tech firms argued that AI should also be used to raise defenders' capabilities as attackers become more powerful.
In the end, the argument is that cybersecurity in the AI era will become a competition between attack AI and defense AI.
The joint letter also drew participation from major security companies such as Cloudflare, CrowdStrike Holdings, Inc., and Palo Alto Networks, along with leading firms in the cloud, finance, and industrial sectors.
[email protected] Lee Gu-soon Reporter