Tech Giants Warn of AI Cyber Threats While Proposing AI-Driven Fixes

OpenAI has organized an open letter signed by more than 100 major technology and information security firms warning that current cybersecurity methods are insufficient to stop increasingly sophisticated AI-enabled attacks. According to reporting from The Register, the signatories include AI developers such as OpenAI, Google, Microsoft, and Anthropic, as well as cloud and software providers like AWS, IBM, Oracle, and Cisco.

Security firms including CrowdStrike, Palo Alto Networks, Fortinet, and Cloudflare also signed the letter, which predicts that AI attacks will become more widespread in the coming months. The group specifically identified internet infrastructure, water treatment plants, and hospitals as being at high risk. The letter notes that these vulnerabilities are compounded by years of unpatched software, weak authentication, and misconfigurations.

To address the crisis, the group proposes a solution centered on more AI. The letter suggests that cheaper AI models handle security at scale while frontier systems are reserved for complex problems. It calls for security vendors to share threat intelligence and for governments to fund AI-powered defenses for essential services. Additionally, frontier model developers are asked to provide training, hands-on support, and "significant funding" to under-resourced critical infrastructure operators.

However, The Register notes that the letter fails to provide specific deadlines, firm commitments, or a dollar amount for the requested funding. The reporting highlights a contradiction in the warning: many of the companies now claiming the "status quo security won't be enough" have spent years selling those very services and tools to their clients.