Abstract:
As artificial intelligence models rapidly improve in their ability to perform cybersecurity tasks, technology industry leaders warn that cyber defenders have limited time left to prepare for the next generation of cyber threats. OpenAI’s recently launched GPT-5.6-Cyber achieves a 95% completion rate for advanced network security requests such as exploit chains, authentication bypasses, and privilege escalations. The company also previously revealed that the upcoming Astra model will have stronger cybersecurity capabilities.
Faced with the imminent advent of more powerful models, many large technology companies around the world believe that network defense agencies must prepare for future threats as soon as possible. OpenAI issued an open letter that day, calling for increased investment in cyber defense globally. More than 100 organizations have signed the letter, including Microsoft, Google, Anthropic, AWS, AMD, Cisco, Cloudflare, CrowdStrike, IBM, Oracle, Palo Alto Networks, Fortinet and other companies.
The companies expect that AI-driven cyberattacks will become more common and more sophisticated in the coming months as model capabilities continue to improve. Hospitals, water treatment facilities, internet infrastructure and other critical services could all be severely impacted, as many of them still rely on legacy systems with unpatched software, weak authentication, over-granted permissions and other security issues.
The open letter believes that more cybersecurity professionals should be given access to highly capable AI models so that they can more quickly discover vulnerabilities, develop repair plans and respond to attacks.
The letter urges organizations to prioritize fixing the highest-risk vulnerabilities and adopt stricter access controls. At the same time, network security vendors should continue to test their products based on cutting-edge AI capabilities and incorporate AI-driven defense mechanisms into existing security tools.
The open letter ends by calling on the industry and government departments to mobilize all technologies, resources and professional capabilities to promote AI with cybersecurity capabilities to prioritize defense teams that protect critical infrastructure; repair the most dangerous weak links, verify the repair effect, and share effective experience so that other institutions can further improve their capabilities on this basis. The letter states that current advances in AI technology should be transformed into long-term security improvements that benefit everyone.
In addition, the government is called upon to provide funding for cybersecurity improvements to hospitals, utilities, local governments, and other under-resourced organizations. Cutting-edge AI companies such as OpenAI and Anthropic should provide model access, funding, training, security tools, and on-site support to defense personnel.
Comments