116 companies and organizations, including OpenAI, Anthropic, Microsoft, Google's parent company Alphabet, and Amazon, have signed a joint letter urging businesses and policymakers to prioritize cybersecurity as the top priority in the AI era, and to "take decisive action" to strengthen defenses. The letter states that in the coming months, as global model capabilities continue to evolve, AI-powered cyberattacks will inevitably become more rampant.

Strengthening defenses on both government and corporate fronts

The joint letter urges governments to accelerate the implementation of the "Trusted Access Program" — allowing specific compliant companies to access advanced models before they are made publicly available. It also calls on all organizations to immediately make network defense a core issue for leadership decisions. Signatories also urge various organizations to comprehensively raise the "security threshold" of defensive tools, and to use low-cost models alongside cutting-edge models when upgrading systems.

The letter also highlights the need for governments to work together: providing dedicated funding for network defense, and focusing on supporting critical infrastructure such as hospitals and water treatment plants, which are frequently attacked but lack resources, effectively lowering the barrier to obtaining protective capabilities. The signatories span a wide range of industries, including cybersecurity companies, financial services firms, semiconductor manufacturers, and cloud service providers.

Notably, Hugging Face, an open-source platform and one of the signatories, has recently been involved in a security breach caused by an uncontrolled OpenAI agent. This attack not only shook the tech and cybersecurity industries, but also brought public scrutiny to the question of whether "AI development is moving too fast."