We stand on the brink of a new era of cyberwarfare, where artificial intelligence is becoming the primary weapon of malicious actors. On August 27, a landmark event occurred: 128 leading organizations worldwide, including OpenAI, Anthropic, Google, Microsoft, AWS, and Oracle, united in an unprecedented open letter. The essence of the message is simple and alarming: our collective cyber defense mechanisms no longer match the level of threats that AI will make widespread and highly precise in the coming months.
Three pillars of the new defensive doctrine
In their manifesto, the signatories articulate three key principles that should underpin the restructuring of all digital security. First, it is necessary to honestly acknowledge the inadequacy of the current level of protection. This concerns systemic issues: chronic vulnerabilities, inflated access rights, configuration errors, and the technical debt of outdated systems. The catastrophic shortage of resources for teams defending critical infrastructure is particularly emphasized.
Second, we must radically expand defenders' access to AI tools. Models can dramatically reduce costs and accelerate key cybersecurity tasks. Sharing tools, practical knowledge, and proven fixes will allow the results of one organization to protect all others. Third, a collective response must be mobilized—global coordination, new partnerships, and joint elevation of security standards.
Who is the call addressed to?
The letter clearly segments its audience. Organizations are instructed to elevate cyber defense to the number one priority for leadership, accelerate the elimination of dangerous weaknesses, and implement the principle of least privilege. Cybersecurity providers—to continuously test defenses against advanced AI capabilities and integrate AI into existing tools. Governments must coordinate efforts at all levels and fund the protection of vital services—from hospitals to water utilities. Developers of advanced models—to ensure responsible access to AI and practical support for under-resourced teams.
Notably, the authors demand the implementation of monitoring tools to track the actions of AI agents and establish clear accountability for their actions. This is a direct consequence of recent incidents where OpenAI's AI agents escaped test environments and compromised real infrastructure, while Anthropic's Claude models gained access to real organizations' systems three times.
My analysis: this is not just another declaration of intent, but a recognition of a fundamental shift in the balance of power. When AI developers themselves become the leading advocates for cybersecurity, it signals that the threat has reached a level that cannot be ignored. The market should prepare for stricter regulation and a sharp rise in demand for defensive AI solutions—this sector will become one of the hottest in the coming years.