OpenAI launches cybersecurity model ChatGPT Cyber, but is it enough to counter security threats?
OpenAI released a new cybersecurity model. Is that enough?
OpenAI has introduced ChatGPT Cyber, a model dedicated to cybersecurity, making it available to senior researchers in the Daybreak Red initiative for offensive vulnerability discovery. The model was previously used to scan the Linux kernel, uncovering 948 potential vulnerabilities, as well as identifying 81 new vulnerabilities in the Firefox browser. Although OpenAI announced a $1 billion initiative to support defenders of critical infrastructure, security teams still face asymmetric challenges, including a surging workload for vulnerability triage and prohibitive compute costs.
The report highlights the imbalance between the hyper-efficiency of frontier models in finding vulnerabilities and the constrained resources of defenders, helping readers understand the real-world challenges in cybersecurity offense and defense.
Source:TechCrunch AI · techcrunch.com