Leading players in the technology industry, including OpenAI, Anthropic, Microsoft, Google, and Amazon Web Services, have joined forces in an unprecedented initiative. A joint open letter, signed by more than a hundred companies from the technology, banking, insurance, and semiconductor manufacturing sectors, warns of an impending wave of cyberattacks using artificial intelligence that will become not only larger in scale but also significantly more sophisticated.
The Essence of the Collective Warning
The authors of the document highlight four key areas for immediate action. First, cybersecurity must be elevated to the level of top company management, rather than remaining the prerogative of IT departments. Second, solution providers need to make AI-driven protection accessible to operators of critical infrastructure. Third, governments should develop coordinated response measures and provide support to hospitals, utility services, and local administrations. Finally, leading AI companies are recommended to open access to their models and fund the training of defense specialists.
"Modern advances in AI are already giving defenders new ways to eliminate weaknesses that have accumulated over the years. If we act decisively, we can use this window of opportunity to make the digital world much safer," the letter emphasizes.
Escalation of Threats: Numbers and Facts
My observations of the market confirm that this is not just a declaration of intent. Analysts at TeamT5 have recorded that Chinese hacker groups have doubled the number of attacks after switching to DeepSeek and other open-source models. Notably, the choice fell not on the most powerful, but on the most cost-effective solutions.
Even more alarming statistics come from Anthropic: an analysis of 832 blocked accounts showed that the share of attackers with medium and high risk levels grew from 33% to 56% over the year. AI now independently handles privilege escalation and lateral movement across networks—tasks that previously required deep technical skills. This means that the barrier to entry into cybercrime has radically lowered.
South Korean researchers at Genians have discovered that the North Korean intelligence group Kimsuky is actively testing local AI tools. An assessment by TRM Labs shows that the use of AI for criminal purposes reached 54 points out of 100 this year, compared to 28 in 2024. In the first half of 2026, 201 hacks of crypto projects were recorded—compared to 83 a year earlier.
My conclusion: we stand on the threshold of a fundamental shift in the balance of power between attackers and defenders. The current surge in AI development can indeed provide more robust protection, but only under the condition of consolidated efforts from the entire digital community. Hesitation here is tantamount to defeat.