As reported by Reuters, on Wednesday, OpenAI issued a warning. It stated that with its performance undergoing rapid enhancements, its soon - to - be - launched new AI model could bring about "high - level" cybersecurity threats. These AI models might possess the ability to autonomously create zero - day remote exploits. Such exploits are capable of launching attacks on highly fortified systems. Moreover, they could play a role in assisting with intricate corporate or industrial intrusion operations, which could have tangible impacts in the real world.

OpenAI mentioned that it is channeling resources into bolstering the models' proficiency in defensive cybersecurity tasks. This involves developing tools that can aid defenders in carrying out code audits and addressing vulnerabilities. In addition, the company will employ measures like access control and infrastructure hardening to mitigate risks. Furthermore, it plans to roll out a program. This program aims to explore the possibility of offering tiered access permissions to qualified users for the purpose of cyber defense. OpenAI will also set up a "Frontier Risk Committee" to work in tandem with security professionals.