More than 100 technology, cybersecurity, infrastructure and financial organizations have backed an open letter calling for a global expansion of AI-assisted cyber defense, with particular attention to critical infrastructure operators that lack the staff or budgets to strengthen their systems.

Published by OpenAI on August 27, the letter is backed by Anthropic, AWS, Google, Microsoft, Cisco, Cloudflare, CrowdStrike, Deutsche Telekom, Equinix, GoDaddy, IBM, Nokia, Oracle and other organizations spanning AI development, Internet infrastructure, telecommunications, security and finance. The signatory list remained open to additional organizations after publication.

The signatories predict a sharp acceleration in AI-enabled cyberattacks. They say such attacks will become “far more widespread and sophisticated” in the coming months as AI models become more capable. That timeframe represents the signatories’ assessment of how quickly the threat may develop, rather than an independently established finding. They identify hospitals, water systems and infrastructure that powers the Internet among the systems at risk, while arguing that the same advances can give defenders greater capacity to find and repair weaknesses.

The letter centers its proposed response on three principles: existing security practices will not be sufficient; more defenders should gain access to cyber-capable AI; and responding to increasingly capable cyber tools requires international cooperation. It points to familiar vulnerabilities—including unpatched software, weak authentication, excessive permissions, misconfigurations and technical debt—as weaknesses that AI-assisted attacks could exploit more effectively.

The recommendations divide responsibility across four groups. Organizations are urged to fix and verify their highest-risk weaknesses and strengthen access controls and defense-in-depth. Cybersecurity companies and technology providers are asked to continuously test defenses against frontier capabilities, make AI-assisted security tools deployable by critical-infrastructure operators, share threat intelligence and provide practical deployment assistance. Governments are asked to fund cyber defense for under-resourced essential services, expand trusted access to defensive AI and strengthen international incident-response coordination.

Frontier AI developers receive a separate set of requests: provide responsible model access, “significant funding,” training and hands-on assistance; develop observability and security tools; make agentic identities traceable; support authorized security testing and private vulnerability disclosure; and share tools, playbooks and threat assessments with governments, security providers and open-source maintainers.

Endorsing the letter does not itself specify what each signatory will contribute. The document sets out recommendations for categories of organizations rather than individual commitments from every company listed. It provides no funding amounts, staffing levels, model-access terms, delivery schedules or common accountability mechanism. TechCrunch also noted that several AI developers supporting the appeal are simultaneously building more capable models while developing programs intended to put frontier AI capabilities to defensive use.