More than 100 companies, including Google, Microsoft, OpenAI and Anthropic, have backed an open letter calling for stronger global cyber defences as advances in artificial intelligence increase the potential scale and sophistication of cyberattacks.
The letter warns that organisations have a limited window to strengthen their security systems, with AI-enabled cyberattacks expected to become more widespread and advanced in the coming months. The companies said existing cybersecurity measures may not be sufficient as AI models become increasingly capable.
The coalition includes companies spanning technology, cybersecurity, financial services and other industries. AWS, Adobe, IBM, Oracle, Cisco, CrowdStrike, Mastercard, Visa, Capital One, Citi, Hugging Face and Deutsche Telekom are among the signatories.
A key concern raised by the companies is the security of critical infrastructure. Hospitals, water treatment facilities and infrastructure supporting internet services were identified among the systems that could face increased risks from AI-assisted attacks.
The letter points to longstanding cybersecurity weaknesses including unpatched software, weak authentication, excessive permissions, system misconfigurations and technical debt in legacy infrastructure. It also argues that security teams, particularly those protecting critical infrastructure, have historically lacked sufficient resources.
However, the companies also see artificial intelligence as part of the defensive response. AI tools can potentially help security teams identify vulnerabilities, automate security tasks and address weaknesses more quickly.
The signatories have called on organisations to make cyber defence an immediate leadership priority and strengthen security standards across the technology they build, purchase and deploy. This includes reviewing AI-generated code and introducing stronger access controls and layered security measures.
Cybersecurity companies and technology providers have also been urged to continuously test their systems against increasingly capable AI models and share threat intelligence and verified fixes where appropriate.
Governments, meanwhile, are being asked to increase coordination and provide funding, technical resources and AI-enabled defensive capabilities to organisations responsible for essential services. The letter also calls on frontier AI companies to provide responsible access to advanced models for cybersecurity defenders, alongside training and support.
The warning comes as AI companies examine how increasingly autonomous models behave in cybersecurity environments. Recent incidents and controlled evaluations involving AI agents have intensified scrutiny around whether advanced models could discover vulnerabilities or perform cyber operations with less human involvement.
The companies argue that the same technological advances creating new security challenges could also give defenders an opportunity to address vulnerabilities that have accumulated over several years.
The collective appeal reflects growing attention around the dual use of advanced AI systems as technology companies, governments and cybersecurity providers assess how rapidly improving models could reshape both offensive and defensive cyber capabilities.
Disclaimer: This article may include information derived from interviews, press releases, public statements, research, company communications and other publicly available or third-party sources. Such material may be summarised, paraphrased or contextualised for journalistic and editorial purposes. All rights in third-party content remain with their respective owners.