"GPT-5.4-Cyber is designed to give verified security professionals access to more powerful and less restrictive AI functionality, necessary as both defenders and attackers use AI."
"What sets GPT-5.4-Cyber apart is its adjusted safety threshold, making it easier for security researchers to analyze vulnerabilities or investigate malware without source code."
"OpenAI is moving away from centralized control over who gains access, aiming to rely on objective verification and usage signals to determine trustworthiness."
OpenAI has introduced GPT-5.4-Cyber, a generative AI model tailored for cybersecurity, emphasizing defensive applications. This model is part of the Trusted Access for Cyber program, allowing verified security professionals to utilize more powerful AI functionalities. The model features a lower safety threshold for trusted users, facilitating vulnerability analysis and malware investigation. OpenAI acknowledges the risks associated with AI technology and links access to identity verification. The initiative builds on Codex Security, which has successfully resolved thousands of vulnerabilities and supports numerous projects with free security scans.
Read at Techzine Global
Unable to calculate read time
Collection
[
|
...
]